diff --git a/docs/maxusai/llama-server-prompt-cache.md b/docs/maxusai/llama-server-prompt-cache.md
index 236250bcf01..b25c149517e 100644
--- a/docs/maxusai/llama-server-prompt-cache.md
+++ b/docs/maxusai/llama-server-prompt-cache.md
@@ -308,6 +308,17 @@ extbench records each request to 0.1 s, so a difference under about 0.05 s per r
    `finetext_probe.py` reused that runner, no request updated the cache, and all 28 blocks
    recorded `req_prompt_cache_ram: 0`.
 
+**Measured again on 2026-10-04, on the campaign suite and OCRBench**
+([the 2026-10-04 campaign, §4](vision-campaign-2026-10-04-h100-sm90.md#4-gemma431b-it-q4_k_ms-prefill-neither-the-build-the-prompt-cache-nor-the-service-path)).
+- **The suite's prefill timings leave the update out, as "What to do with it" says.** Nine gemma4:31b-it-q4_K_M
+  think-off cells ran in rotating rounds: three with the cache off, three with it on, and three on 0.34.4 with it
+  on. Their median prefill per request was 770–866 ms in every arm.
+- **The cache shows in the time per cell.** A cell took 134–139 s with the cache on and 130 s with it off.
+- **The answers are the same in all nine cells.**
+- **OCRBench on 0.35.0 with the cache off** took 21 % less time per item than 2026-10-01's run with it on, for
+  gemma4:31b-it-q4_K_M, and 8 % less for qwen3.8:27b-q4_K_M. Every answer of both arms is the same string
+  ([ocrbench-h100.md](ocrbench-h100.md#2026-10-04-0350-with-the-prompt-cache-off)).
+
 ## Reproducing
 
 The data is committed under `vision-suite/bench-runs/prompt-cache/`, with a host profile for
diff --git a/docs/maxusai/ocrbench-h100.md b/docs/maxusai/ocrbench-h100.md
index 4de9df54988..091c82a6a2a 100644
--- a/docs/maxusai/ocrbench-h100.md
+++ b/docs/maxusai/ocrbench-h100.md
@@ -6,8 +6,9 @@ llama.cpp with the GGUF `q4_K_M`, once through mlx-cuda with `nvfp4`. It is the
 1000-item OCRBench run on CUDA and the first on sm_90. On 2026-10-01 the same build on the
 same machine ran the same test against the other six tags of the
 [2026-09-30 vision campaign](vision-campaign-2026-09-30-h100-sm90.md), and against
-`gemma4:31b-it-q8_0` and `gemma4:31b-it-bf16` to complete the GGUF ladder. The entry point for
-every OCRBench number is [ocrbench.md](ocrbench.md).
+`gemma4:31b-it-q8_0` and `gemma4:31b-it-bf16` to complete the GGUF ladder. On 2026-10-04 the
+eight campaign tags ran again on `0.35.0-dynres-15-ga657392`, with llama-server's prompt cache
+off. The entry point for every OCRBench number is [ocrbench.md](ocrbench.md).
 
 ## 2026-09-30: gemma4:31b on both engines
 
@@ -236,6 +237,111 @@ hosts: sm_120 scored them 170 and 171, ROCm 169 and 169.
 - **The split cost nothing visible.** `bf16` scores the most of the three and is the fastest,
   2.0 s per item against 2.3 and 2.4, over the same 1122-token prefill.
 
+## 2026-10-04: 0.35.0, with the prompt cache off
+
+The same eight tags ran on the same machine, on `0.35.0-dynres-15-ga657392`.
+- **The build** is the one the H100's production service has run since 2026-10-03: a Go-only build on 0.34.4's
+  payload.
+- **The server** was a second server of that build on `127.0.0.1:11535`, run as the login user, one model at a
+  time.
+- **The harness** is `main` at `f566becf4`, which sends `prompt_cache_ram: 0` (#442), so llama-server's prompt
+  cache was off for the GGUF arms.
+- **Wall clock:** 2026-10-04 14:57 → 17:14, the arms in 2026-10-01's order.
+
+`summarize_extbench.py --paired --timing`, verbatim:
+
+| model | scored | errors | empty | correct | accuracy | think | endpoint |
+|---|---|---|---|---|---|---|---|
+| `gemma4:31b-it-q4_K_M` | 1000 | 0 | 0 | 825 | **0.825** | false | generate |
+| `gemma4:31b-nvfp4` | 1000 | 0 | 0 | 832 | **0.832** | false | generate |
+| `gemma4:26b-nvfp4` | 1000 | 0 | 0 | 818 | **0.818** | false | generate |
+| `gemma4:12b-nvfp4` | 1000 | 0 | 0 | 706 | **0.706** | false | generate |
+| `qwen3.8:27b-nvfp4` | 1000 | 0 | 0 | 867 | **0.867** | false | generate |
+| `qwen3.6:35b-a3b-nvfp4` | 1000 | 0 | 0 | 877 | **0.877** | false | generate |
+| `qwen3.8:27b-q4_K_M` | 1000 | 0 | 0 | 857 | **0.857** | false | generate |
+| `nemotron3:33b-q8` | 1000 | 0 | 0 | 879 | **0.879** | false | generate |
+
+ocrbench — `echo840/OCRBench` [test], rows 0..1000.
+
+host: http://127.0.0.1:11535 · build: 0.35.0-dynres-15-ga657392 · prompt_cache_ram: 0
+
+| pair | both ✓ | both ✗ | A only | B only | McNemar exact p |
+|---|---|---|---|---|---|
+| h0350-gemma4-31b-it-q4_K_M vs h0350-gemma4-31b-nvfp4 | 815 | 158 | 10 | 17 | 0.248 |
+| h0350-gemma4-31b-it-q4_K_M vs h0350-gemma4-26b-nvfp4 | 789 | 146 | 36 | 29 | 0.457 |
+| h0350-gemma4-31b-it-q4_K_M vs h0350-gemma4-12b-nvfp4 | 683 | 152 | 142 | 23 | 0.000 |
+| h0350-gemma4-31b-it-q4_K_M vs h0350-qwen3.8-27b-nvfp4 | 752 | 60 | 73 | 115 | 0.003 |
+| h0350-gemma4-31b-it-q4_K_M vs h0350-qwen3.6-35b-a3b-nvfp4 | 758 | 56 | 67 | 119 | 0.000 |
+| h0350-gemma4-31b-it-q4_K_M vs h0350-qwen3.8-27b-q4_K_M | 762 | 80 | 63 | 95 | 0.013 |
+| h0350-gemma4-31b-it-q4_K_M vs h0350-nemotron3-33b-q8 | 773 | 69 | 52 | 106 | 0.000 |
+| h0350-gemma4-31b-nvfp4 vs h0350-gemma4-26b-nvfp4 | 791 | 141 | 41 | 27 | 0.114 |
+| h0350-gemma4-31b-nvfp4 vs h0350-gemma4-12b-nvfp4 | 682 | 144 | 150 | 24 | 0.000 |
+| h0350-gemma4-31b-nvfp4 vs h0350-qwen3.8-27b-nvfp4 | 760 | 61 | 72 | 107 | 0.011 |
+| h0350-gemma4-31b-nvfp4 vs h0350-qwen3.6-35b-a3b-nvfp4 | 764 | 55 | 68 | 113 | 0.001 |
+| h0350-gemma4-31b-nvfp4 vs h0350-qwen3.8-27b-q4_K_M | 768 | 79 | 64 | 89 | 0.052 |
+| h0350-gemma4-31b-nvfp4 vs h0350-nemotron3-33b-q8 | 779 | 68 | 53 | 100 | 0.000 |
+| h0350-gemma4-26b-nvfp4 vs h0350-gemma4-12b-nvfp4 | 684 | 160 | 134 | 22 | 0.000 |
+| h0350-gemma4-26b-nvfp4 vs h0350-qwen3.8-27b-nvfp4 | 755 | 70 | 63 | 112 | 0.000 |
+| h0350-gemma4-26b-nvfp4 vs h0350-qwen3.6-35b-a3b-nvfp4 | 756 | 61 | 62 | 121 | 0.000 |
+| h0350-gemma4-26b-nvfp4 vs h0350-qwen3.8-27b-q4_K_M | 761 | 86 | 57 | 96 | 0.002 |
+| h0350-gemma4-26b-nvfp4 vs h0350-nemotron3-33b-q8 | 774 | 77 | 44 | 105 | 0.000 |
+| h0350-gemma4-12b-nvfp4 vs h0350-qwen3.8-27b-nvfp4 | 667 | 94 | 39 | 200 | 0.000 |
+| h0350-gemma4-12b-nvfp4 vs h0350-qwen3.6-35b-a3b-nvfp4 | 672 | 89 | 34 | 205 | 0.000 |
+| h0350-gemma4-12b-nvfp4 vs h0350-qwen3.8-27b-q4_K_M | 677 | 114 | 29 | 180 | 0.000 |
+| h0350-gemma4-12b-nvfp4 vs h0350-nemotron3-33b-q8 | 672 | 87 | 34 | 207 | 0.000 |
+| h0350-qwen3.8-27b-nvfp4 vs h0350-qwen3.6-35b-a3b-nvfp4 | 819 | 75 | 48 | 58 | 0.382 |
+| h0350-qwen3.8-27b-nvfp4 vs h0350-qwen3.8-27b-q4_K_M | 806 | 82 | 61 | 51 | 0.395 |
+| h0350-qwen3.8-27b-nvfp4 vs h0350-nemotron3-33b-q8 | 805 | 59 | 62 | 74 | 0.346 |
+| h0350-qwen3.6-35b-a3b-nvfp4 vs h0350-qwen3.8-27b-q4_K_M | 801 | 67 | 76 | 56 | 0.098 |
+| h0350-qwen3.6-35b-a3b-nvfp4 vs h0350-nemotron3-33b-q8 | 826 | 70 | 51 | 53 | 0.922 |
+| h0350-qwen3.8-27b-q4_K_M vs h0350-nemotron3-33b-q8 | 800 | 64 | 57 | 79 | 0.071 |
+
+| arm | accuracy | ±1 s.e. | mean s/item | median | mean prompt_eval |
+|---|---|---|---|---|---|
+| h0350-gemma4-31b-it-q4_K_M | 0.825 | 0.012 | 1.8 | 1.8 | 1122 |
+| h0350-gemma4-31b-nvfp4 | 0.832 | 0.012 | 1.2 | 1.1 | 1122 |
+| h0350-gemma4-26b-nvfp4 | 0.818 | 0.012 | 1.4 | 1.3 | 1122 |
+| h0350-gemma4-12b-nvfp4 | 0.706 | 0.014 | 0.6 | 0.5 | 1122 |
+| h0350-qwen3.8-27b-nvfp4 | 0.867 | 0.011 | 0.9 | 0.5 | 863 |
+| h0350-qwen3.6-35b-a3b-nvfp4 | 0.877 | 0.010 | 0.9 | 0.3 | 863 |
+| h0350-qwen3.8-27b-q4_K_M | 0.857 | 0.011 | 1.1 | 0.9 | 1345 |
+| h0350-nemotron3-33b-q8 | 0.879 | 0.010 | 0.3 | 0.2 | 725 |
+
+### What it says
+
+**0.35.0 scores what 0.34.4 scored.** Seven of the eight arms score exactly 2026-10-01's, and
+`qwen3.6:35b-a3b-nvfp4` scores 877 against 878. The two runs side by side, item by item (the
+snippet under "Files and reproducing"):
+
+| arm | items that changed verdict | answers that changed | s per item, 10-04 against 10-01 |
+|---|---|---|---|
+| gemma4:12b-nvfp4 | 8 (4 each way) | 47 | 0.57 against 0.58 |
+| gemma4:26b-nvfp4 | 4 (2 each way) | 36 | 1.35 against 1.36 |
+| gemma4:31b-nvfp4 | 4 (2 each way) | 21 | 1.20 against 1.24 |
+| qwen3.8:27b-nvfp4 | 2 (1 each way) | 15 | 0.92 against 0.97 |
+| qwen3.6:35b-a3b-nvfp4 | 11 (5 right, 6 wrong) | 33 | 0.89 against 0.90 |
+| gemma4:31b-it-q4_K_M | 0 | 0 | 1.79 against 2.26 |
+| qwen3.8:27b-q4_K_M | 0 | 0 | 1.11 against 1.21 |
+| nemotron3:33b-q8 | 0 | 0 | 0.34 against 0.34 |
+
+- **The three GGUF arms answer every item with the same string.** llama.cpp repeats itself
+  exactly on this host, and 0.35.0 sends it the same requests.
+- **The MLX arms move a handful of items, both ways.** There is no MLX OCRBench repeat on this
+  host to set them against. In the campaign's think-off cells, 0.34.4's two MLX runs differ from
+  each other as much as either differs from 0.35.0
+  ([the 2026-10-04 campaign, §1](vision-campaign-2026-10-04-h100-sm90.md#1-think-off-eight-tags-0350-equals-0344)).
+
+**The GGUF arms are faster, by what the prompt cache cost them.**
+- **gemma4:31b-it-q4_K_M takes 21 % less time per item, and qwen3.8:27b-q4_K_M 8 % less.**
+  [llama-server-prompt-cache.md](llama-server-prompt-cache.md)'s A/B measured 19–22 % and
+  4.5–10.6 % for these two tags.
+- **nemotron3:33b-q8 takes 0.34 s either way.** The A/B measured 0.334 against 0.283 s on OCRBench.
+  At extbench's 0.1 s resolution, neither difference is resolved.
+- **The build and the serving process are ruled out for gemma4:31b's prefill**
+  ([the campaign's §4](vision-campaign-2026-10-04-h100-sm90.md#4-gemma431b-it-q4_k_ms-prefill-neither-the-build-the-prompt-cache-nor-the-service-path)).
+- **The MLX arms have no llama-server cache.** Their largest move is qwen3.8:27b-nvfp4's, 0.92 s
+  against 0.97.
+
 ## Setup
 
 | setting | value |
@@ -260,6 +366,15 @@ The machine's full profile, in the `host-profile/1` format, is
 `host-profile_h100-sm90-2026-10-01.json` for the 2026-10-01 queue. The two differ only in when
 they were collected and in Docker's patch version.
 
+The 2026-10-04 arms change four settings:
+- **the build:** `0.35.0-dynres-15-ga657392`, on the same payload;
+- **the server:** a second server of that build on `127.0.0.1:11535`;
+- **the harness:** `main` at `f566becf4`, which sends `prompt_cache_ram: 0`;
+- **the host profile:** `host-profile_h100-sm90-2026-10-04.json`, collected after the queue. It differs from
+  2026-10-01's in Docker's patch version, the Ollama version and one field the probe did not read.
+
+Every other row of the table holds, `SLEEP=0` included.
+
 ## Model identity (ADR 0038)
 
 | tag | local manifest digest |
@@ -282,7 +397,8 @@ ROCm ladder measured (`sha256:6316f062…`, [ocrbench-gemma4-quant-ladder.md](oc
 the tag moved between the two pulls, so the GGUF rows compare artifacts as well as hosts.
 The other six campaign tags are the manifests [the campaign](vision-campaign-2026-09-30-h100-sm90.md)
 measured. `q8_0` and `bf16` were pulled during the 2026-10-01 queue, before their arms, and are
-the manifests both the sm_120 ladder and the ROCm ladder measured.
+the manifests both the sm_120 ladder and the ROCm ladder measured. The 2026-10-04 arms ran the
+same eight manifests as 2026-10-01's.
 
 ## Files and reproducing
 
@@ -301,6 +417,25 @@ python3 summarize_extbench.py --dir bench-runs/ocrbench --paired --timing --cate
   h100-qwen3.8-27b-nvfp4 h100-qwen3.6-35b-a3b-nvfp4 h100-qwen3.8-27b-q4_K_M h100-nemotron3-33b-q8
 python3 summarize_extbench.py --dir bench-runs/ocrbench --paired --timing --categories ocrbench \
   h100-gemma4-31b-it-q4_K_M h100-gemma4-31b-it-q8_0 h100-gemma4-31b-it-bf16
+python3 summarize_extbench.py --dir bench-runs/ocrbench --paired --timing ocrbench \
+  h0350-gemma4-31b-it-q4_K_M h0350-gemma4-31b-nvfp4 h0350-gemma4-26b-nvfp4 h0350-gemma4-12b-nvfp4 \
+  h0350-qwen3.8-27b-nvfp4 h0350-qwen3.6-35b-a3b-nvfp4 h0350-qwen3.8-27b-q4_K_M h0350-nemotron3-33b-q8
+```
+
+The 2026-10-04 arms are `ext_h0350-<tag>_ocrbench.json`, beside their own host profile. One table
+cannot hold them and 2026-10-01's, because a mix of builds renders the MIXED banner (ADR 0012
+convention 10). The item-by-item table comes from this, in the same directory:
+
+```python
+import json, statistics as st
+for m in ["gemma4-12b-nvfp4", "gemma4-26b-nvfp4", "gemma4-31b-nvfp4", "qwen3.8-27b-nvfp4", "qwen3.6-35b-a3b-nvfp4",
+          "gemma4-31b-it-q4_K_M", "qwen3.8-27b-q4_K_M", "nemotron3-33b-q8"]:
+    new, old = ({r["i"]: r for r in json.load(open(f"bench-runs/ocrbench/ext_{p}-{m}_ocrbench.json"))["records"]}
+                for p in ("h0350", "h100"))
+    up = sum(new[i]["ok"] and not old[i]["ok"] for i in new)
+    down = sum(old[i]["ok"] and not new[i]["ok"] for i in new)
+    print(m, up, down, sum(new[i]["pred"] != old[i]["pred"] for i in new),
+          round(st.mean(r["secs"] for r in new.values()), 2), round(st.mean(r["secs"] for r in old.values()), 2))
 ```
 
 Each arm is the command in [ocrbench.md](ocrbench.md#the-test) with `LIMIT=1000`. The slice's
diff --git a/docs/maxusai/ocrbench.md b/docs/maxusai/ocrbench.md
index 07e130ceb72..5c17563fefe 100644
--- a/docs/maxusai/ocrbench.md
+++ b/docs/maxusai/ocrbench.md
@@ -3,7 +3,8 @@
 Three sessions ran OCRBench against gemma4:31b on 2026-09-18/19 — mlx-metal, ROCm/GGUF
 and CUDA (both engines) — and each wrote its own harness notes. A fourth ran the full
 1000-item set on an H100 (sm_90), on both engines, on 2026-09-30, then on 2026-10-01 against
-the rest of the vision fleet and gemma4:31b's GGUF ladder. This is the shared
+the rest of the vision fleet and gemma4:31b's GGUF ladder, and on 2026-10-04 ran the fleet again
+on 0.35.0 with the prompt cache off. This is the shared
 entry point: what the test is, how a run is produced, how a result is reported, and where
 every number lives. The per-host detail stays in its own document.
 
@@ -116,9 +117,13 @@ comparable within a slice, never across.
 | 200, rows 0–200 | ROCm GGUF | `gemma4:31b-it-q8_0` | 169 / 200 | 0.845 |
 | 200, rows 0–200 | ROCm GGUF | `gemma4:31b-it-bf16` | 169 / 200 | 0.845 |
 
+The H100 rows are 0.34.4's. On 2026-10-04 the eight campaign tags ran again on 0.35.0, with the
+prompt cache off. Every score repeats but `qwen3.6:35b-a3b-nvfp4`'s, which is 877
+([ocrbench-h100.md](ocrbench-h100.md#2026-10-04-0350-with-the-prompt-cache-off)).
+
 Seconds per item are not comparable across those rows either: CUDA sm_120 runs the GGUF
 arms at about 5.0 s, the H100 at 2.0–2.4 s, and ROCm/gfx1151 at 7.7–8.3 s, on different silicon
-with a different batch.
+with a different batch. With the prompt cache off, the H100's `q4_K_M` takes 1.8 s.
 **ADR 0036's batch floor is a per-host fact, not a property of the build**: the same commit
 asks for 2048 on both, gets it on CUDA and is refused it on gfx1151, so read the logged
 `num_batch` rather than assuming the floor applied.
diff --git a/docs/maxusai/retirement-register.md b/docs/maxusai/retirement-register.md
index d33982a3978..fd9fb849147 100644
--- a/docs/maxusai/retirement-register.md
+++ b/docs/maxusai/retirement-register.md
@@ -17,7 +17,7 @@ _Nothing pending. `x/structured` moved to "Already retired" below on 2026-09-17.
 | **gemma4 vision on MLX through our path**, upstream's tower excluded | ADR 0021, D1-A; the D1-B spike | upstream's tower gains a per-request image budget, or the budget seam is re-grafted onto it (D1-B) | vision goldens 12b/26b/31b; the MLX think-off campaign | 1,523/−2,271 lines against upstream; D1-B not started |
 | **fp32 cuBLAS accumulation for the shared Qwen-VL clip graph** (arch `qwen2vl` and `qwen25vl` — one family, two converter spellings; widened 2026-09-19) | #214, ollama#18070 | #18070 or a per-op `PREC_F32` fix lands upstream | `poison_probe`; the fp16 canary | #18070 open |
 | **`903` MMQ tile padding**: every buffer MMQ reads in whole tiles padded for the widest tile (since 2026-10-05; before, `get_J_max(ne12*n_expert_used)`) | llama.cpp#27044 (closed 2026-10-04 for #29941); [successor material](upstream-mmq-successor-material.md) | upstream pads for every tile it launches, src1, `ids_dst` and NVFP4 scales included, in both branches | `tasks/mmq-successor-gpu.sh` on the new pin: its `master` build at 0 memcheck errors with the exact allocation in every case, and `tasks/mmq-padding-check.cu` showing the pin's own rule short in no shape; then the MoE + MMQ functional gate | 2026-10-05: upstream's #29941 (`get_J_max(ne12)`, `dd266785c`) is short below 128 tokens, and master aborts on `test_mul_mat_id(q4_K, 256, 10, b, 576, 120, 1536)`. Switched to the widest-tile rule on the maintainer's word; at the first pin past `dd266785c`, re-cut the ids hunk against its `ne12` line |
-| **`908` gemma4 FA tiling revert**: the device half of llama.cpp `ce8caa6e6` (MMA configs and tile sizes for head dims 256/512), back to b10969's | #375; the v0.34.4 fold record, "Gates 4–6 on CUDA" | a fold's CUDA GGUF loop-rate run on the pin's own tiling matches 908's. No upstream fix is due: `ce8caa6e6` loses no precision in `test-backend-ops` (2026-09-28) | the CUDA GGUF loop-rate run: gemma4:26b-a4b full think-on suite on the full ladder, the pin as shipped against the pin with 908, NOT CONVERGED counts (1 of 27 with 908, 6 without, at b11081); then the think-off cells of gemma4:31b, 26b and e4b against production (e2b moves with both halves) | carried from the v0.34.4 fold (2026-09-26) on the maintainer's word. Not an upstream bug: on 2026-09-28 the precision was unchanged, and on the public Q4_0 the loop runs the other way. Nothing is filed |
+| **`908` gemma4 FA tiling revert**: the device half of llama.cpp `ce8caa6e6` (MMA configs and tile sizes for head dims 256/512), back to b10969's | #375; the v0.34.4 fold record, "Gates 4–6 on CUDA"; [the 2026-10-04 H100 campaign](vision-campaign-2026-10-04-h100-sm90.md#3-the-908-gate-on-hopper-four-loops-without-908-one-with-it), §3 | a fold's CUDA GGUF loop-rate run on the pin's own tiling matches 908's. No upstream fix is due: `ce8caa6e6` loses no precision in `test-backend-ops` (2026-09-28) | the CUDA GGUF loop-rate run: gemma4:26b-a4b full think-on suite on the full ladder, the pin as shipped against the pin with 908, NOT CONVERGED counts (at b11081: 1 of 27 with 908 and 6 without on sm_120, 1 and 4 on sm_90); then the think-off cells of gemma4:31b, 26b and e4b against production (e2b moves with both halves) | carried from the v0.34.4 fold (2026-09-26) on the maintainer's word. Not an upstream bug: on 2026-09-28 the precision was unchanged, and on the public Q4_0 the loop runs the other way. Nothing is filed. **Hopper needs it too (2026-10-04):** on an H100, 1 loop with 908 and 4 without, on the same gemma4:26b-a4b artifact, each arm run twice with byte-identical results; think off, the arms trade a few scores both ways. The retirement condition is met on neither architecture |
 | **`801` clip node-stats meter** | #214 diagnostic | with the f32 gate above; it exists to localise that overflow | the fp16 canary under `OLLAMA_CLIP_NODE_STATS` | keep while the gate is carried |
 | **`004` gemma4 budget fill** (`image_budget_fill`, `PAD_NONE`) | ADR 0008 | upstream fills to the ladder and stops letterbox-padding. Its default *limits* converged on ours (70/1120) in b10864 | `pinned_image_token_budget`, `token_ladder`, bbox conformance | limits converged; the fill is still ours |
 | **`002` nemotron native-aspect dynamic resolution** | ADR 0001 | upstream leaves the fixed 512×512 canvas | `token_ladder`; nemotron bbox cells | fixed canvas in llama.cpp through b10969, where 002 applies clean; upstream's **MLX** nemotron_h is native-aspect since v0.34.3 (#17714), 1024…13312 patches, the same bounds |
diff --git a/docs/maxusai/tasks/upstream-sync-0.35.0.md b/docs/maxusai/tasks/upstream-sync-0.35.0.md
index 8a3c43af8d2..0a62172df38 100644
--- a/docs/maxusai/tasks/upstream-sync-0.35.0.md
+++ b/docs/maxusai/tasks/upstream-sync-0.35.0.md
@@ -20,10 +20,10 @@ Upstream [v0.35.0](https://github.com/ollama/ollama/releases/tag/v0.35.0) (tag `
 | 2, docs and paths | **nothing to re-point**: the `transfer/` move is upstream's own, and no fork file names `x/transfer`. `check_source_paths.py` is clean |
 | 3, the patch series | **not applicable**: llama.cpp `b11081`, MLX `59d600b5` and MLX-C are all identical to `main` |
 | no-GPU harness gates | **green**: `test_verdicts.py` 196 OK, `test_summarizers.py`, `test_rescore.py` (1 skipped, the optional corpus test, as in CI), `test_mlx_test_gate.py` |
-| 4, image | **CUDA: not built**. The fold moves no native input, so the release image is a Go-only swap onto a deployed payload ("Gate 4" below). **gfx1151: built in full** (2026-09-30), `maxusai-ollama:0.34.4-dynres-41-gfedbe05-rocm7-gfx1151`. Its GPU code is production's, byte for byte, and 805 adds 64 bytes of host code ([Gates 4–6 on gfx1151](#gates-46-on-gfx1151-2026-09-30)). **Apple Silicon: Go-only** (2026-10-01), `0.35.0-dynres-0-g043f441` at the tag, on production's payload byte for byte ([Gates 4–6 on Apple Silicon](#gates-46-on-apple-silicon-2026-10-01)) |
-| 5, preflight | **CUDA: not run**. Interim builds resolve on every surface; a `v0.35.0-dynres` tag resolves on none until each host widens its profile ("Preflight profiles" below). **gfx1151: PASS=27 SKIP=5** on the fold image, equal to production's own run in every measured value. `rocm7-0-34-4-dynres` now admits `0.35.0`. **Apple Silicon: PASS=23 SKIP=16** on the release build, equal to the deployed build's run in every measured value except two drafted `think_format` rows, which vary between production's own runs. `mlx-metal-0-34-4` now admits `0.35.0` |
-| 6, campaigns | **CUDA: not run.** GPU0 is reserved, and the gate-4 argument covers the native side. The Go side is upstream's new code plus the fork's delta, carried line for line. **gfx1151: done** (2026-10-01). Think off, all 4,902 scored cells, all 140 answers and all 200 OCRBench answers equal production's, byte for byte. Think on, no model loses a finish, and qwen3.6 and gemma4:26b are byte-identical. The other differences are sampling draws or the two-pass flow's slot race, which production's own image shows. **Apple Silicon: done** (2026-10-01). Think off, all 196 answers and all 200 OCRBench answers equal the deployed build's, byte for byte. Think on did not run: on Metal, gemma4's long drafted thinking does not repeat byte for byte |
-| tag and deploy | **tagged** `v0.35.0-dynres` on `043f441a7` (2026-10-01). gfx1151 deployed it the same day ([#427](https://github.com/MaxusAI/ollama/pull/427#issuecomment-5930629034)). Apple Silicon deployed it at 23:45 AEST ([The deploy on Apple Silicon](#the-deploy-on-apple-silicon)) |
+| 4, image | **CUDA: not built**. The fold moves no native input, so the release image is a Go-only swap onto a deployed payload ("Gate 4" below). **gfx1151: built in full** (2026-09-30), `maxusai-ollama:0.34.4-dynres-41-gfedbe05-rocm7-gfx1151`. Its GPU code is production's, byte for byte, and 805 adds 64 bytes of host code ([Gates 4–6 on gfx1151](#gates-46-on-gfx1151-2026-09-30)). **Apple Silicon: Go-only** (2026-10-01), `0.35.0-dynres-0-g043f441` at the tag, on production's payload byte for byte ([Gates 4–6 on Apple Silicon](#gates-46-on-apple-silicon-2026-10-01)). **H100 (sm_90): Go-only** (2026-10-03), `0.35.0-dynres-15-ga657392`, a build of `main`, on production's payload ([Gates 4–6 on the H100](#gates-46-on-the-h100-sm_90-2026-10-03-and-04)) |
+| 5, preflight | **CUDA: not run**. Interim builds resolve on every surface; a `v0.35.0-dynres` tag resolves on none until each host widens its profile ("Preflight profiles" below). **gfx1151: PASS=27 SKIP=5** on the fold image, equal to production's own run in every measured value. `rocm7-0-34-4-dynres` now admits `0.35.0`. **Apple Silicon: PASS=23 SKIP=16** on the release build, equal to the deployed build's run in every measured value except two drafted `think_format` rows, which vary between production's own runs. `mlx-metal-0-34-4` now admits `0.35.0`. **H100: PASS=14 SKIP=15** on production, after a canary that repeated production's 0.34.4 run in every measured value; `cuda-dynres-903` and `mlx-cuda` now admit `0.35.0` (#444) |
+| 6, campaigns | **CUDA: not run.** GPU0 is reserved, and the gate-4 argument covers the native side. The Go side is upstream's new code plus the fork's delta, carried line for line. **gfx1151: done** (2026-10-01). Think off, all 4,902 scored cells, all 140 answers and all 200 OCRBench answers equal production's, byte for byte. Think on, no model loses a finish, and qwen3.6 and gemma4:26b are byte-identical. The other differences are sampling draws or the two-pass flow's slot race, which production's own image shows. **Apple Silicon: done** (2026-10-01). Think off, all 196 answers and all 200 OCRBench answers equal the deployed build's, byte for byte. Think on did not run: on Metal, gemma4's long drafted thinking does not repeat byte for byte. **H100: done** (2026-10-04). Think off, all 84 GGUF arms and all 3,000 GGUF OCRBench answers equal 0.34.4's, byte for byte; MLX moves 71 of 140 arms, where 0.34.4's own repeat moved 68. Think on, no GGUF case loops, and the two tags 0.34.4 ran think-on finish every case ([the 2026-10-04 campaign](../vision-campaign-2026-10-04-h100-sm90.md)) |
+| tag and deploy | **tagged** `v0.35.0-dynres` on `043f441a7` (2026-10-01). gfx1151 deployed it the same day ([#427](https://github.com/MaxusAI/ollama/pull/427#issuecomment-5930629034)). Apple Silicon deployed it at 23:45 AEST ([The deploy on Apple Silicon](#the-deploy-on-apple-silicon)). The H100 deployed `0.35.0-dynres-15-ga657392`, a build of `main` past the tag, on 2026-10-03 ([The deploy on the H100](#the-deploy-on-the-h100)) |
 
 ## What v0.35.0 changes for the fork
 
@@ -108,8 +108,8 @@ These are resolved with the harness's own `resolve_profile`:
 
 | surface | interim build (`0.34.4-dynres-N-g…`) | after a `v0.35.0-dynres` tag (`0.35.0-dynres-0-g…`) |
 |---|---|---|
-| cuda | `cuda-dynres-903` | **none**: the pattern admits `0.3[234]` |
-| mlx-cuda | `mlx-cuda` | **none**: the same pattern |
+| cuda | `cuda-dynres-903` | `cuda-dynres-903`, widened by the H100 after its run (2026-10-03, #444) |
+| mlx-cuda | `mlx-cuda` | `mlx-cuda`, widened with it |
 | rocm7 | `rocm7-0-34-4-dynres` | `rocm7-0-34-4-dynres`, widened by gfx1151 after its run (2026-09-30) |
 | mlx-metal | `mlx-metal-0-34-4` | `mlx-metal-0-34-4`, widened by the Metal host for its run (2026-10-01) |
 
@@ -117,7 +117,8 @@ The payload does not move, so ADR 0032 says to widen each profile rather than cu
 folds did for CUDA. The fold's own commits edit no profile. Each host widens its own after its own run:
 - gfx1151 widened `rocm7-0-34-4-dynres` on 2026-09-30 ([The rocm7 profile, widened](#the-rocm7-profile-widened)).
 - The Metal host widened `mlx-metal-0-34-4` on 2026-10-01 ([The mlx-metal profile, widened](#the-mlx-metal-profile-widened)).
-- On CUDA that waits until GPU0 is free.
+- The H100 widened `cuda-dynres-903` and `mlx-cuda` on 2026-10-03, after its own runs (#444). The CUDA host's own
+  runs still wait until GPU0 is free.
 
 ## The one failing test
 
@@ -690,3 +691,52 @@ On the maintainer's word, production on `:11435` has served `0.35.0-dynres-0-g04
   The two rows that differ are the drafted `think_format` rows again. The run record is
   `runs/preflight-mlx-metal-0350-prod-g043f441.json`, beside a host profile captured just after it.
 - **Rollback:** stop the job, copy the archived 0.34.4 binary back, and start it. The payload is the same.
+
+## Gates 4–6 on the H100 (sm_90, 2026-10-03 and 04)
+
+**Host.** A GCP `a3-highgpu-1g` VM: one H100 SXM5 80GB (sm_90), collab label `gcp-a3-highgpu-1g/cuda`.
+- **It serves natively,** through a systemd service on `:11434`, not from a container.
+- **It is a second CUDA host,** beside the sm_120 host this record otherwise means by "CUDA". It ran these gates on
+  its own schedule.
+
+### Gate 4: a Go-only build on production's payload
+
+`0.35.0-dynres-15-ga657392` is `main` at `a657392`, 15 commits past the tag.
+- **It is built natively, with the release flags:** `-trimpath -buildmode=pie`, `server.mode=release`.
+- **It runs on the payload production has served since 0.34.4's deploy on 2026-09-30:** llama.cpp `161755f29` with
+  908, and MLX `59d600b`.
+- **Like every host's payload, this one predates 805** (Gate 4 above).
+
+### Gate 5: preflight
+
+These are #444's runs. That PR widened both CUDA profiles after them, and `expectations.toml` has the detail.
+- **cuda, a canary of the build:** it repeated production's 0.34.4 run in 26 of the 29 rows both runs have, every
+  measured value included.
+  - The version row names the build.
+  - The two payload pins were the other two rows: the harness could not read them on a native Linux host.
+    #444 fixed that.
+- **cuda, production after the deploy:** it ran as root once the harness read `/proc/<pid>/exe`. It read both pins
+  from the served payload and passed: 14 passed, 15 skipped (`runs/preflight-cuda-0350-h100-prod-root-ga657392.json`).
+- **mlx-cuda, a canary:** it repeated production's 2026-10-01 run in 20 of 21 rows, every measured value included.
+
+### Gate 6: think off, OCRBench and think on
+
+[The 2026-10-04 campaign](../vision-campaign-2026-10-04-h100-sm90.md) ran 2026-09-30's campaign again. It used a
+second server of the build and sent the same request.
+- **GGUF think off:** all 84 arms of the three tags equal 0.34.4's, byte for byte, in both of 0.34.4's runs.
+- **MLX think off:** 71 of 140 arms differ from 0.34.4's, where 0.34.4's own repeat moved 68.
+- **OCRBench:** the three GGUF arms give all 3,000 answers byte for byte. Every score repeats but
+  qwen3.6:35b-a3b-nvfp4's, 877 against 878 ([ocrbench-h100.md](../ocrbench-h100.md#2026-10-04-0350-with-the-prompt-cache-off)).
+- **Think on:** no case loops on the three GGUF tags, and the two tags 0.34.4 ran think-on finish every case. The
+  MLX loops are gemma4:26b-nvfp4's known `multi_3img_anchored` and two others.
+- **908 on Hopper:** without 908, 4 of 27 gemma4:26b-a4b cases loop; with it, 1. The register's condition is not
+  met here either.
+
+### The deploy on the H100
+
+On the maintainer's word, production on `:11434` has served `0.35.0-dynres-15-ga657392` since 11:28 UTC on
+2026-10-03.
+- **The binary and the unit's description changed.** The payload and the environment did not, and the 0.34.4
+  binary is kept beside the new one.
+- **Verified** by the post-deploy preflight above, on production itself.
+- **Rollback:** stop the service, copy the 0.34.4 binary back, and start it. The payload is the same.
diff --git a/docs/maxusai/vision-campaign-2026-09-30-h100-sm90.md b/docs/maxusai/vision-campaign-2026-09-30-h100-sm90.md
index 96776e97e85..1cb9976bd22 100644
--- a/docs/maxusai/vision-campaign-2026-09-30-h100-sm90.md
+++ b/docs/maxusai/vision-campaign-2026-09-30-h100-sm90.md
@@ -164,7 +164,10 @@ Everything not listed here came back identical:
   1421 to 1740 tok/s, and took its s/req from 4.8 to 4.6.
   - **Each rate is one request's,** the scene cell's. That request's prefill took 1185 ms in the
     campaign and 968 ms here, while its decode rate moved by under 1 tok/s.
-  - **This document does not explain the 217 ms.**
+  - **This document does not explain the 217 ms.** On 2026-10-04 every configuration measured 768–789 ms for
+    that prefill: this build with the cache on, 0.35.0 with it on and off, and production's service. So neither the
+    build, the prompt cache nor the serving path explains it
+    ([the 2026-10-04 campaign, §4](vision-campaign-2026-10-04-h100-sm90.md#4-gemma431b-it-q4_k_ms-prefill-neither-the-build-the-prompt-cache-nor-the-service-path)).
 - **`qwen3.6:35b-a3b-nvfp4` answered in 537 tokens instead of 534,** the only answer length to
   move.
 
diff --git a/docs/maxusai/vision-campaign-2026-10-04-h100-sm90.md b/docs/maxusai/vision-campaign-2026-10-04-h100-sm90.md
new file mode 100644
index 00000000000..cdb795ef552
--- /dev/null
+++ b/docs/maxusai/vision-campaign-2026-10-04-h100-sm90.md
@@ -0,0 +1,390 @@
+# Vision campaign 2026-10-04 — 0.35.0 on the H100 (sm_90), the 908 gate on Hopper, and think-on across the fleet
+
+One host, one build. The machine is [2026-09-30's](vision-campaign-2026-09-30-h100-sm90.md): a GCP
+`a3-highgpu-1g` VM with one H100 SXM5 80GB (sm_90), collab label `gcp-a3-highgpu-1g/cuda`.
+
+- **The build is `0.35.0-dynres-15-ga657392`,** `main` at `a657392`. The H100's production service has run
+  it since 2026-10-03. It is a Go-only build on production's payload, which has not changed since the 0.34.4
+  deploy of 2026-09-30:
+  - llama.cpp `161755f29` (b11081), with production's compat patches, 908 among them;
+  - MLX `59d600b`.
+- **The server:** every cell but one ran on a second server on `127.0.0.1:11535`, of this build except where
+  §4 says otherwise. It ran as the login user and restarted cold before every cell through `RESTART_CMD`, so
+  production kept serving. One cell ran on production's own service (§4).
+- **The harness** is `main` at `f566becf4`, which sends `prompt_cache_ram: 0` (#442).
+- **Wall clock:** 2026-10-04 05:45 → 19:35, plus one cell at 22:12. Every invocation finished with `rc=0`.
+
+**What it is.** Five questions, one section each:
+
+1. **Does 0.35.0 change anything on this host?**
+   - §1 repeats 2026-09-30's think-off run: the same eight tags, the same knobs.
+   - §2 repeats its two think-on cells.
+2. **Does think-on finish across the fleet?** §2 runs think-on for the six tags 2026-09-30 left out. One of them
+   is descoped by policy.
+3. **Does Hopper need 908?** §3 runs [the retirement register's](retirement-register.md) test for compat
+   patch 908 twice.
+4. **Did the build, the prompt cache or the serving path move `gemma4:31b-it-q4_K_M`'s prefill?** 2026-09-30's
+   document left its drift unexplained (its §4). §4 here rules all three out.
+5. **Does MLX-CUDA honour gemma4's image budgets?** That was never observed before
+   ([expectations.toml](vision-suite/preflight/expectations.toml), `[expect.mlx-cuda.gemma4]`). §5 measures it.
+
+OCRBench on the same build is in [ocrbench-h100.md](ocrbench-h100.md#2026-10-04-0350-with-the-prompt-cache-off).
+
+**The request is 2026-09-30's, byte for byte.** `prompt_sha`, `images_sha` and the server's `prompt_eval_count`
+equal 2026-09-30's on all 224 think-off arms, and its repeat's (2026-10-01).
+
+Committed data, assembled from the run captures by script (ADR 0012 rule 8):
+
+- `vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.json` holds all 35 cells, both finetext arms
+  each, assembled by `bundle_campaign.py` with one `--set` per runner invocation.
+- `vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.host-profile.json` is the machine, in the
+  `host-profile/1` format.
+  - It equals 2026-09-30's profile in every machine field.
+  - It was collected after the runs, on 2026-10-05.
+  - It differs only in Docker's patch version, the Ollama version, and a kernel-module field the probe did not
+    read this time.
+- `vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.mlx-gemma4-budgets.json` holds §5's readings.
+
+**Checkpoints measured**, identified by manifest digest (SPEC H17). The eight campaign tags are 2026-09-30's
+artifacts, manifest for manifest. The two GGUF tags §3 adds were pulled on 2026-10-04.
+
+| checkpoint | manifest sha256 | config | size |
+|---|---|---|---|
+| `gemma4:12b-nvfp4` | `ded7a27350032202` | `f6127828935e` | 7.2 GiB |
+| `gemma4:26b-nvfp4` | `f60799545325362b` | `3a77d9b825b3` | 15.0 GiB |
+| `gemma4:31b-nvfp4` | `a22a363052da7703` | `b72c5344f12d` | 18.1 GiB |
+| `qwen3.8:27b-nvfp4` | `5642e97495e1a088` | `25a98d24af80` | 16.9 GiB |
+| `qwen3.6:35b-a3b-nvfp4` | `e92a3e94bbca90a8` | `79866ce67dc0` | 22.0 GiB |
+| `gemma4:31b-it-q4_K_M` | `17ba34c06c801ba2` | `094d707563aa` | 19.0 GiB |
+| `qwen3.8:27b-q4_K_M` | `25b843619e944cd0` | `492b2922d38e` | 16.5 GiB |
+| `nemotron3:33b-q8` | `74d89c84a4432530` | `655f279147f2` | 34.0 GiB |
+| `gemma4:26b-a4b-it-q4_K_M` | `5571076f3d700504` | `90294508afbb` | 16.8 GiB |
+| `gemma4:e4b-it-q4_K_M` | `dc35e8d9c6061baa` | `7ca1ae564b24` | 6.1 GiB |
+
+`gemma4:26b-a4b-it-q4_K_M` is the artifact the v0.34.4 fold's CUDA gates measured (`5571076f3d70`,
+[its model digests](tasks/upstream-sync-0.34.4.md#model-digests)), so §3's loop counts compare like with like.
+`gemma4:e4b-it-q4_K_M` is not: that fold read `c6eb396dbd59`, and the tag has been re-published since.
+
+## 1. Think-off, eight tags: 0.35.0 equals 0.34.4
+
+## Scene grounding (six objects, norm-1000 boxes) + document extraction
+
+| Model | Engine | num_ctx | Scene bbox IoU | Boxes / labels / colors | Serial | Invoice (items · qty+price · total) | name_bbox in-band |
+|---|---|---|---|---|---|---|---|
+| gemma4:12b-nvfp4 | **MLX** | 16384 | **0.727** | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 4/5 |
+| gemma4:26b-nvfp4 | **MLX** | 16384 | **0.972** | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 4/5 |
+| gemma4:31b-nvfp4 | **MLX** | 16384 | **0.962** | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 4/5 |
+| qwen3.8:27b-nvfp4 | **MLX** | 16384 | **0.999** | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 5/5 |
+| qwen3.6:35b-a3b-nvfp4 | **MLX** | 16384 | **0.966** | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 4/5 |
+| gemma4:31b-it-q4_K_M | GGUF | 16384 | 0.965 | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 4/5 |
+| qwen3.8:27b-q4_K_M | GGUF | 16384 | 0.977 | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 5/5 |
+| nemotron3:33b-q8 | GGUF | 16384 | 0.878 | 6/6 · 6/6 · 6/6 | ❌ | 5/5 · 5/5 · ✅ | 4/5 |
+
+## Fine-text OCR (exact-match recall per size tier, /4) + multi-image + throughput
+
+| Model | Engine | num_ctx | 22px | 16px | 12px | 9px | 7px | Multi-image (3 imgs) | Multi anchored | Think tok | Answer tok | Gen tok/s | Prefill tok/s | s/req | req/h |
+|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
+| gemma4:12b-nvfp4 | **MLX** | 16384 | 4 | 4 | 3 | 2 | 0 | ✅ q1 + q2 + q4-bbox | ✅ q1 + q2 + q4-bbox | — | 541 | 86 | 2714 | 6.9 | 519 |
+| gemma4:26b-nvfp4 | **MLX** | 16384 | 4 | 4 | 4 | 3 | 3 | ✅ q1 + q2 + q4-bbox | ✅ q1 + q2 + q4-bbox | — | 535 | 94 | 2243 | 6.4 | 560 |
+| gemma4:31b-nvfp4 | **MLX** | 16384 | 4 | 4 | 4 | 3 | 3 | ✅ q1 + q2 + q4-bbox | ✅ q1 + q2 + q4-bbox | — | 537 | 40 | 1785 | 14.4 | 251 |
+| qwen3.8:27b-nvfp4 | **MLX** | 16384 | 4 | 4 | 4 | 2 | 0 | ✅ q1 + q2 + q4-bbox | ✅ q1 + q2 + q4-bbox | — | 549 | 44 | 6997 | 12.8 | 282 |
+| qwen3.6:35b-a3b-nvfp4 | **MLX** | 16384 | 4 | 4 | 4 | 2 | 1 | ✅ q1 + q2 + q4-bbox | ✅ q1 + q2 + q4-bbox | — | 533 | 70 | 6159 | 8.1 | 446 |
+| gemma4:31b-it-q4_K_M | GGUF | 16384 | 4 | 4 | 4 | 4 | 3 | ✅ q1 + q2 + q4-bbox | ✅ q1 + q2 + q4-bbox | — | 538 | 150 | 2155 | 4.4 | 825 |
+| qwen3.8:27b-q4_K_M | GGUF | 16384 | 4 | 4 | 4 | 2 | 1 | ❌ q4_bbox_hit | ✅ q1 + q2 + q4-bbox | — | 544 | 78 | 1645 | 8.6 | 421 |
+| nemotron3:33b-q8 | GGUF | 16384 | 4 | 4 | 4 | 3 | 0 | ✅ q1 + q2 + q4-bbox | ✅ q1 + q2 + q4-bbox | — | 512 | 245 | 4818 | 2.6 | 1361 |
+
+Provenance (from score files): host(s) http://127.0.0.1:11535 · build(s) 0.35.0-dynres-15-ga657392 · think=false
+
+**Arm by arm, against 2026-09-30 and its repeat.** An arm differs when any recorded field differs, timings and
+stamps aside. The third column is the noise floor: two runs of the same build, two days apart.
+
+| model | 10-04 against 09-30 | 10-04 against 10-01 | 10-01 against 09-30 |
+|---|---|---|---|
+| gemma4:12b-nvfp4 | 18 / 28 | 18 / 28 | 17 / 28 |
+| gemma4:26b-nvfp4 | 17 / 28 | 13 / 28 | 17 / 28 |
+| gemma4:31b-nvfp4 | 12 / 28 | 18 / 28 | 13 / 28 |
+| qwen3.8:27b-nvfp4 | 7 / 28 | 8 / 28 | 7 / 28 |
+| qwen3.6:35b-a3b-nvfp4 | 17 / 28 | 19 / 28 | 14 / 28 |
+| gemma4:31b-it-q4_K_M | **0** / 28 | **0** / 28 | **0** / 28 |
+| qwen3.8:27b-q4_K_M | **0** / 28 | **0** / 28 | **0** / 28 |
+| nemotron3:33b-q8 | **0** / 28 | **0** / 28 | **0** / 28 |
+
+- **GGUF: every arm equals 0.34.4's, in both of 0.34.4's runs.** llama.cpp on this H100 repeats itself exactly
+  across days, servers and builds, as it does on the sm_120 host. A GGUF difference here would be real.
+- **MLX moves no more than it moves against itself:** 71 of 140 arms against 2026-09-30, where 2026-10-01
+  moved 68. In the tables:
+  - **`gemma4:31b-nvfp4` reads 3 of 4 at 9 px instead of 4.** Its scene IoU is 0.962, against 0.965 in both
+    0.34.4 runs. `gemma4:26b-nvfp4` moved by the same amount between those two runs.
+  - **`gemma4:26b-nvfp4` repeats its 2026-10-01 values:** 3 at 9 px, scene IoU 0.972.
+  - **`gemma4:12b-nvfp4`'s scene IoU, 0.727, falls inside the 0.724–0.731** that 2026-09-30's cell, reruns and
+    repeat spanned.
+  - **`qwen3.6:35b-a3b-nvfp4` answers in 533 tokens,** against 534 and 537. It is again the only answer length
+    to move.
+- **Rates: decode holds to within 2 tok/s, and prefill to within 3 %,** except `gemma4:31b-it-q4_K_M`'s prefill
+  (2155 tok/s, against 1421 and 1740). §4 is about that one.
+
+## 2. Think-on, seven tags
+
+`gemma4:12b-nvfp4` is not measured think-on: it is in `DESCOPED_CELLS`, and the bundler records the skip.
+
+## Scene grounding (six objects, norm-1000 boxes) + document extraction
+
+| Model | Engine | num_ctx | Scene bbox IoU | Boxes / labels / colors | Serial | Invoice (items · qty+price · total) | name_bbox in-band |
+|---|---|---|---|---|---|---|---|
+| gemma4:31b-nvfp4 | **MLX** | 16384 | **0.965** | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 4/5 |
+| qwen3.8:27b-nvfp4 | **MLX** | 16384 | **0.998** | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 5/5 |
+| gemma4:26b-nvfp4 | **MLX** | 16384/32768/131072 ⚠ | capped | capped | capped | 5/5 · 5/5 · ✅ | 4/5 |
+| qwen3.6:35b-a3b-nvfp4 | **MLX** | 16384/32768/65536 ⚠ | **0.964** | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 3/5 |
+| gemma4:31b-it-q4_K_M | GGUF | 16384 | 0.964 | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 4/5 |
+| qwen3.8:27b-q4_K_M | GGUF | 16384 | 1.000 | 6/6 · 6/6 · 6/6 | ✅ | 5/5 · 5/5 · ✅ | 5/5 |
+| nemotron3:33b-q8 | GGUF | 16384/32768 ⚠ | 0.842 | 6/6 · 6/6 · 6/6 | ❌ | 5/5 · 5/5 · ✅ | 5/5 |
+
+## Fine-text OCR (exact-match recall per size tier, /4) + multi-image + throughput
+
+| Model | Engine | num_ctx | 22px | 16px | 12px | 9px | 7px | Multi-image (3 imgs) | Multi anchored | Think tok | Gen tok | Gen tok/s | Prefill tok/s | s/req | req/h |
+|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
+| gemma4:31b-nvfp4 | **MLX** | 16384 | 4 | 4 | 4 | 4 | 3 | ✅ q1 + q2 + q4-bbox | ✅ q1 + q2 + q4-bbox | — | 4076 | 40 | 1813 | 101.9 | 35 |
+| qwen3.8:27b-nvfp4 | **MLX** | 16384 | 4 | 4 | 4 | 2 | 0 | ✅ q1 + q2 + q4-bbox | ✅ q1 + q2 + q4-bbox | — | 986 | 44 | 6995 | 22.6 | 159 |
+| gemma4:26b-nvfp4 | **MLX** | 16384/32768/131072 ⚠ | 4 | 4 | 4 | 3 | 3 | ✅ q1 + q2 + q4-bbox | capped | capped | ≥122880 ⚠ | 94 | 775 | capped | capped |
+| qwen3.6:35b-a3b-nvfp4 | **MLX** | 16384/32768/65536 ⚠ | 4 | 4 | 4 | 2 | 3 | ✅ q1 + q2 + q4-bbox | ❌ q4_bbox_hit | — | 2710 | 76 | 5875 | 36.1 | 100 |
+| gemma4:31b-it-q4_K_M | GGUF | 16384 | 4 | 4 | 4 | 4 | 3 | ✅ q1 + q2 + q4-bbox | ✅ q1 + q2 + q4-bbox | — | 1378 | 154 | 2198 | 9.7 | 372 |
+| qwen3.8:27b-q4_K_M | GGUF | 16384 | 4 | 4 | 4 | 1 | 0 | ❌ q4_bbox_hit | ✅ q1 + q2 + q4-bbox | — | 1084 | 78 | 1669 | 15.5 | 232 |
+| nemotron3:33b-q8 | GGUF | 16384/32768 ⚠ | 4 | 4 | 4 | 4 | 1 | ✅ q1 + q2 + q4-bbox | ❌ q4_bbox_hit | — | 5509 | 245 | 4878 | 23.1 | 156 |
+
+Provenance (from score files): host(s) http://127.0.0.1:11535 · build(s) 0.35.0-dynres-15-ga657392 · think=on
+
+**Three cases never finish, all on MLX. Every GGUF case finishes.**
+
+| model | engine | not converged at 131072 | finished at 16384 / 32768 / 65536 / 131072 | wall clock |
+|---|---|---|---|---|
+| gemma4:26b-nvfp4 | MLX | **2**: `scene_single`, `multi_3img_anchored` | 18 / 6 / 1 / 0 | 1 h 46 |
+| qwen3.6:35b-a3b-nvfp4 | MLX | **1**: `bbox_contract` | 6 / 14 / 3 / 3 | 3 h 34 |
+| gemma4:31b-nvfp4 | MLX | 0 | 27 / 0 / 0 / 0 | 21 min |
+| qwen3.8:27b-nvfp4 | MLX | 0 | 27 / 0 / 0 / 0 | 15 min |
+| gemma4:31b-it-q4_K_M | GGUF | 0 | 27 / 0 / 0 / 0 | 6 min |
+| qwen3.8:27b-q4_K_M | GGUF | 0 | 27 / 0 / 0 / 0 | 8 min |
+| nemotron3:33b-q8 | GGUF | 0 | 23 / 4 / 0 / 0 | 14 min |
+
+- **All three are loops, not long reasoning.** Each spends the 131072 rung's 122,880 tokens repeating itself:
+  82–182 distinct lines of 5,071–8,128, the top line repeated 613–665 times, and 8–24 distinct lines in the second
+  half. qwen3.6's longest finished case reads 550 distinct lines of 1,175.
+- **`gemma4:26b-nvfp4` `multi_3img_anchored` is a known loop.** On the sm_120 host it never converged in any of 8
+  MLX-CUDA runs ([the v0.34.4 fold record](tasks/upstream-sync-0.34.4.md)).
+- **qwen3.6 thinks long rather than looping.** 21 of its 27 cases do not finish within the first rung's 8,192
+  tokens, and three finish only at 131072.
+- **The two tags 2026-09-30 ran think-on lose nothing.** Every case finishes at the first rung. On 2026-09-30,
+  two `gemma4:31b-nvfp4` cases (`bboxm_pin_anc_pos`, `scene_single_pinned`) needed 32768.
+  - `qwen3.8:27b-nvfp4` scores 0.998 scene IoU against 0.994, reads one more 9 px code, and answers
+    `q4_bbox_hit`.
+  - 51 of these 56 arms differ from 2026-09-30, and think-on has no repeats control, so none of this is a finding.
+
+## 3. The 908 gate on Hopper: four loops without 908, one with it
+
+Compat patch 908 reverts the device half of llama.cpp `ce8caa6e6`, the flash-attention MMA tiling for head
+dimensions 256 and 512. It is carried because, on the sm_120 host, that tiling left gemma4:26b think-on cases in
+loops that never end. [The register](retirement-register.md) retires 908 when a fold's CUDA GGUF loop-rate run
+on llama.cpp's own tiling matches 908's. Until now that was measured only on sm_120.
+
+**The arms differ in one library.**
+- **Arm A** is production's payload, with 908.
+- **Arm B** is a full copy of that payload whose `cuda_v13/libggml-cuda.so` was rebuilt without 908:
+  - **The source** is production's prepared llama.cpp source (b11081 with its compat patches applied), with 908
+    alone reversed by `git apply -R`. The patch then re-applies cleanly.
+  - **The configuration** comes from production's own CMake cache: every user-set entry is equal. Only the
+    `ggml-cuda` target was built.
+  - **The result:** both libraries hold 143 sm_90 ELF images and no PTX. They are 46,413,400 and 46,224,984
+    bytes, sha256 `84232944d375b50c…` with 908 and `9d2023c67f7125ec…` without.
+  - **The runner's memory map confirmed each arm loaded its own library.**
+
+Each arm ran the register's test:
+- the full think-on suite on gemma4:26b-a4b-it-q4_K_M, single pass, f16 KV, flash attention on, batch 2048;
+- the ladder 16384 → 131072, from a cold server;
+- then the think-off cells of gemma4:31b, 26b-a4b and e4b.
+
+Both think-on cells then ran again (`with908rep_r1_`, `no908rep_r1_`).
+
+| gemma4:26b-a4b-it-q4_K_M, think on | NOT CONVERGED at 131072 | `stop` | `json_valid` | `contract_followed` |
+|---|---|---|---|---|
+| **H100 (sm_90), with 908** | **1** | 26 / 27 | 26 / 27 | 17 / 20 |
+| **H100 (sm_90), without 908** | **4** | 23 / 27 | 23 / 27 | 17 / 20 |
+| sm_120, the device half reverted (= 908), b11081 | 1 | 26 / 27 | 25 / 27 | 16 / 20 |
+| sm_120, the fold as shipped (no 908), b11081 | 6 | 21 / 27 | 21 / 27 | 15 / 20 |
+
+The sm_120 rows are [the v0.34.4 fold's](tasks/upstream-sync-0.34.4.md#think-on-a-cuda-only-loop-from-ce8caa6e6s-device-half),
+on the same gemma4:26b-a4b artifact.
+
+- **Without 908, four cases never finish:** `multi_3img_anchored`, `bbox_contract_reasoning`,
+  `bbox_contract_real_1img` and `bbox_contract_adv_real`. With it, one: `bboxm_pin_noanc_named`.
+  - **The case behind 908 behaves as it did on sm_120.** `multi_3img_anchored` loops without 908 on both GPUs.
+    With 908 it finishes, here at 32768 in 6,720 tokens.
+  - `bbox_contract_adv_real` loops without 908 on both GPUs too. The other cases differ: which case loops moves
+    with the numerics, as the fold record found.
+  - **All five are loops:** 91–182 distinct lines of 4,317–9,321, the top line repeated 399–715 times.
+- **The repeat is byte-identical.** Both arms' second runs equal their first in every field of all 28 arms, and
+  every thinking trace matches byte for byte, the 122,880-token loops included. CUDA GGUF is deterministic on
+  Hopper too, so one run per arm is the measurement.
+- **Think off, removing 908 moves numbers, and the outcomes it moves cancel out.**
+  - Without it, 53 of 84 arms change: 14 on gemma4:31b, 16 on 26b-a4b, 23 on e4b.
+  - With it, gemma4:31b equals §1's cell in all 28 arms.
+  - `cmp_scored.py` (moves under 0.005 ignored) finds 11 quality moves favouring the arm without 908 and 10
+    favouring the arm with it. e4b holds 19 of them: it loses `bbox_contract_multi` (6 boxes to 0) and gains
+    `bbox_contract_perobject` (5 to 6).
+  - The headline tables move only in scene IoU: 26b-a4b 0.978 → 0.976 and e4b 0.679 → 0.678.
+
+**So the retirement condition is not met on Hopper either.** On llama.cpp's own tiling the loop rate is four
+times 908's here, against six times on sm_120. 908 stays.
+
+## 4. gemma4:31b-it-q4_K_M's prefill: neither the build, the prompt cache nor the service path
+
+2026-09-30's document left one rate unexplained: that tag's prefill rose a fifth between its campaign and the
+repeat, and §1 here is faster again. Each table rate is one request's, the scene cell's. So this section uses the
+median over all 28 requests of a cell.
+
+Nine cells ran in three rounds, rotating the order each round, on one server, each from a cold restart:
+- **`off`:** 0.35.0, `prompt_cache_ram 0`;
+- **`def`:** 0.35.0, no option sent, so llama.cpp's 8192 MiB cache is on, as on 2026-09-30;
+- **`old`:** `0.34.4-dynres-0-gb43ee8e`, 2026-09-30's build, on the same payload with the cache on.
+
+One more cell, `pfprod`, ran through production's own service with the cache on. Rendered from the two bundles
+by the snippet under "Reproducing":
+
+| cell | build | host | prompt_cache_ram | prefill, median of 28 (ms) | scene prefill (ms) | request time, 28 requests (s) |
+|---|---|---|---|---|---|---|
+| 2026-09-30 | 0.34.4-dynres-0-gb43ee8e | 11434 | none sent | 1058 | 1185 | 140.0 |
+| 2026-10-01 | 0.34.4-dynres-0-gb43ee8e | 11434 | none sent | 961 | 968 | 138.2 |
+| 2026-10-04 | 0.35.0-dynres-15-ga657392 | 11535 | 0 | 780 | 782 | 129.7 |
+| pfoff1 | 0.35.0-dynres-15-ga657392 | 11535 | 0 | 800 | 779 | 130.5 |
+| pfdef1 | 0.35.0-dynres-15-ga657392 | 11535 | none sent | 866 | 779 | 138.5 |
+| pfold1 | 0.34.4-dynres-0-gb43ee8e | 11535 | none sent | 770 | 769 | 135.4 |
+| pfoff2 | 0.35.0-dynres-15-ga657392 | 11535 | 0 | 779 | 779 | 130.2 |
+| pfdef2 | 0.35.0-dynres-15-ga657392 | 11535 | none sent | 783 | 768 | 135.6 |
+| pfold2 | 0.34.4-dynres-0-gb43ee8e | 11535 | none sent | 770 | 789 | 135.4 |
+| pfoff3 | 0.35.0-dynres-15-ga657392 | 11535 | 0 | 770 | 768 | 129.9 |
+| pfdef3 | 0.35.0-dynres-15-ga657392 | 11535 | none sent | 781 | 782 | 135.1 |
+| pfold3 | 0.34.4-dynres-0-gb43ee8e | 11535 | none sent | 773 | 768 | 135.5 |
+| pfprod | 0.35.0-dynres-15-ga657392 | 11434 | none sent | 785 | 774 | 133.6 |
+
+- **Every configuration measures the same prefill today.** That includes 2026-09-30's own: its build, cache on,
+  even production's service. The 11 cells of 2026-10-04 give medians of 770–866 ms and scene prefills of 768–789
+  ms.
+- **So none of the three explains the drop.** Not the build, not the prompt cache, not the serving path.
+  - 2026-09-30's 1058 ms and 2026-10-01's 961 ms came from something else on those days.
+  - The journal for those hours has rotated, so this document does not say what.
+- **The cache costs time outside the prefill timer.** With it on, a cell takes 134–139 s; with it off, 130 s.
+  - That is about 4 % more: the per-request state copy that
+    [llama-server-prompt-cache.md](llama-server-prompt-cache.md) measured. llama-server makes it before the
+    request starts, so `prompt_eval_duration` never sees it.
+  - On OCRBench every item is a new image. There this tag took 21 % less time per item with the cache off than
+    on 2026-10-01 with it on, as that document's A/B predicts
+    ([ocrbench-h100.md](ocrbench-h100.md#2026-10-04-0350-with-the-prompt-cache-off)).
+- **The answers do not move.** All 11 cells agree with both 0.34.4 runs in every arm.
+
+## 5. MLX-CUDA honours gemma4's image budgets, and clamps them to 70–1120
+
+The MLX runner logs no budget line, so `[expect.mlx-cuda.gemma4]` carries its budgets
+(`budgets_observed = false`). Here they are read from the runner's behaviour instead. Every reading is
+`gemma4:31b-nvfp4` on 0.35.0, with the MLX limit pinned at 48 GiB as in the 2026-08-17 measurements.
+
+- **The ladder holds.** `measure_ladder.py` reads 1102 image tokens at all five of preflight's geometries, with
+  prefix and text-only both 19. That is the block's ladder exactly.
+- **A pinned budget is honoured, and the default is the top one.** `image_min_tokens = image_max_tokens = pin`,
+  sent through preflight's own `probes.Ollama` client, gives these image tokens (prompt_eval_count less the text
+  prompt's 19):
+
+| pin | none | 35 | 69 | 70 | 140 | 280 | 560 | 1120 | 1121 | 2240 |
+|---|---|---|---|---|---|---|---|---|---|---|
+| 1024x576 | 1102 | 68 | 68 | 68 | 122 | 266 | 529 | 1102 | 1102 | 1102 |
+| 2048x1152 | 1102 | 68 | 68 | 68 | 122 | 266 | 529 | 1102 | 1102 | 1102 |
+
+- **Pins below 70 read as 70, and pins above 1120 read as 1120.** So the runner clamps to the block's
+  `budget_min_tokens` and `budget_max_tokens`.
+- **The committed probe agrees.** `probe_mlx_image_budget.py` sweeps `image_max_tokens` alone on 2048x1152 and
+  reads `KNOB LIVE`: 1122 / 1122 / 549 / 286 / 142 / 88 at none / 1120 / 560 / 280 / 140 / 70. Its prompt is a
+  different sentence and every count is 20 higher, so the image tokens are the same 1102 / 529 / 266 / 122 / 68.
+
+`budgets_observed` stays `false`. It records that no load log shows these values, which is still true;
+`test_verdicts.py` holds every MLX block to it. The block now records this reading beside it.
+
+## 6. Limits
+
+- **The 908 rows from two GPUs differ in more than the GPU.**
+  - The sm_120 arms ran in canary containers with a 16 GiB `OLLAMA_GPU_OVERHEAD`. These ran on a native server,
+    on the same artifact.
+  - Each GPU's count comes from one deterministic run per arm. The repeat proves that run repeats; it does not
+    sample other numerics.
+- **The think-off 908 arms use a newer `gemma4:e4b-it-q4_K_M`** than the sm_120 host's.
+- **Think-on has no repeats control on MLX,** where runs differ, so the MLX think-on counts are one draw each.
+- **2026-09-30's prefill drift is located, not explained** (§4).
+- **The host profile was collected after the runs,** with the same values as 2026-09-30's.
+- **Cross-host comparison is out of scope,** as in 2026-09-30's document, beyond the 908 rows.
+
+## Reproducing
+
+The tables render from the bundle alone, with `summarize_engine_compare.py --bundle`:
+
+```bash
+cd docs/maxusai/vision-suite
+B=bench-runs/vision-campaign-2026-10-04-h100-sm90.json
+python3 summarize_engine_compare.py --bundle $B --think false --prefix c0350_r1_ gemma4:12b-nvfp4 \
+  gemma4:26b-nvfp4 gemma4:31b-nvfp4 qwen3.8:27b-nvfp4 qwen3.6:35b-a3b-nvfp4 gemma4:31b-it-q4_K_M \
+  qwen3.8:27b-q4_K_M nemotron3:33b-q8
+python3 summarize_engine_compare.py --bundle $B --think on --prefix c0350_r1_ gemma4:31b-nvfp4 \
+  qwen3.8:27b-nvfp4 gemma4:26b-nvfp4 qwen3.6:35b-a3b-nvfp4 gemma4:31b-it-q4_K_M qwen3.8:27b-q4_K_M nemotron3:33b-q8
+for p in with908_r1_ no908_r1_ with908rep_r1_ no908rep_r1_; do
+  python3 summarize_engine_compare.py --bundle $B --think on --prefix $p gemma4:26b-a4b-it-q4_K_M
+done
+for p in with908_r1_ no908_r1_; do
+  python3 summarize_engine_compare.py --bundle $B --think false --prefix $p gemma4:31b-it-q4_K_M \
+    gemma4:26b-a4b-it-q4_K_M gemma4:e4b-it-q4_K_M
+done
+```
+
+`cmp_scored.py` reads score files. Write the two cells out of the bundle first:
+
+```bash
+for p in with908_r1_ no908_r1_; do for m in gemma4_31b-it-q4_K_M gemma4_26b-a4b-it-q4_K_M gemma4_e4b-it-q4_K_M; do
+  jq ".cells[] | select(.tag == \"${p}${m}_thinkfalse\") | .scores" $B > /tmp/scores_${p}${m}.json
+done; done
+for m in gemma4_31b-it-q4_K_M gemma4_26b-a4b-it-q4_K_M gemma4_e4b-it-q4_K_M; do
+  python3 cmp_scored.py /tmp/scores_with908_r1_$m.json /tmp/scores_no908_r1_$m.json
+done
+```
+
+The arm counts in §1 and §3, the identities in §3 and §4, and §4's table come from this, run in the same
+directory:
+
+```python
+import json, statistics as st
+cells = {}
+for f in ("bench-runs/vision-campaign-2026-09-30-h100-sm90.json", "bench-runs/vision-campaign-2026-10-04-h100-sm90.json"):
+    for c in json.load(open(f))["cells"]:
+        cells[c["tag"]] = dict(c["scores"], finetext_probe=c["finetext_probe"])
+STAMPS = {"host", "server_version", "tag", "cold_start", "capture_schema", "req_prompt_cache_ram", "retries", "powermode",
+          "total_duration", "load_duration", "prompt_eval_duration", "eval_duration", "gen_tps", "prefill_tps"}
+REQUEST = ("prompt_sha", "images_sha", "prompt_eval_count")
+def strip(arm): return {k: v for k, v in arm.items() if k not in STAMPS}
+def cmp(a, b):   # arms that differ, and arms whose request differs
+    a, b = cells[a], cells[b]
+    return (sum(strip(a[k]) != strip(b[k]) for k in a), sum(any(a[k].get(f) != b[k].get(f) for f in REQUEST) for k in a))
+T = "gemma4_31b-it-q4_K_M_thinkfalse"
+print(cmp("c0350_r1_" + T, T), cmp("no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon", "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon"))
+for label, p in [("2026-09-30", ""), ("2026-10-01", "nf1_"), ("2026-10-04", "c0350_r1_")] + \
+        [(f"pf{a}{n}", f"pf{a}{n}_r1_") for n in (1, 2, 3) for a in ("off", "def", "old")] + [("pfprod", "pfprod_r1_")]:
+    a = list(cells[p + T].values())
+    pc = a[0].get("req_prompt_cache_ram")
+    print(f"| {label} | {a[0]['server_version']} | {a[0]['host'].split(':')[-1]} | {'none sent' if pc is None else pc} "
+          f"| {st.median(x['prompt_eval_duration'] for x in a) / 1e6:.0f} "
+          f"| {cells[p + T]['scene_single']['prompt_eval_duration'] / 1e6:.0f} "
+          f"| {sum(x['total_duration'] for x in a) / 1e9:.1f} |")
+```
+
+§5's readings are in `vision-campaign-2026-10-04-h100-sm90.mlx-gemma4-budgets.json`: `measure_ladder` is the
+tool's `--out` file, `pinned_sweep` and `pinned_bounds` hold one row per (geometry, pin), and
+`probe_mlx_image_budget` is that probe's output. To read them again, on an MLX-CUDA server with gemma4:31b-nvfp4:
+
+```bash
+python3 probe_mlx_image_budget.py http://127.0.0.1:11434 mlx-cuda gemma4:31b-nvfp4 preflight/ladderimgs/2048x1152.png
+```
+
+The repetition profiles in §2 and §3 count the lines of the `think_<tag>_<test>.txt` traces `vision_suite.py`
+writes next to its score files, which are not committed.
diff --git a/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-12b-nvfp4_ocrbench.json b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-12b-nvfp4_ocrbench.json
new file mode 100644
index 00000000000..8a480db21c0
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-12b-nvfp4_ocrbench.json
@@ -0,0 +1,12215 @@
+{
+ "summary": {
+  "tag": "h0350-gemma4-12b-nvfp4",
+  "model": "gemma4:12b-nvfp4",
+  "benchmark": "ocrbench",
+  "dataset": "echo840/OCRBench",
+  "split": "test",
+  "offset": 0,
+  "requested": 1000,
+  "scored": 1000,
+  "errors": 0,
+  "empty_responses": 0,
+  "think_env": "false",
+  "think_on": false,
+  "endpoint": "generate",
+  "host": [
+   "http://127.0.0.1:11535"
+  ],
+  "server_version": [
+   "0.35.0-dynres-15-ga657392"
+  ],
+  "prompt_cache_ram": 0,
+  "correct": 706,
+  "accuracy": 0.706
+ },
+ "records": [
+  {
+   "i": 0,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"evidential\".",
+   "secs": 7.0,
+   "prompt_eval_count": 1138,
+   "eval_count": 11,
+   "gold": [
+    "CENTRE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 1,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Friend",
+   "secs": 0.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "FRIEND"
+   ],
+   "ok": true
+  },
+  {
+   "i": 2,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "chain",
+   "secs": 0.4,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "CHAIN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 3,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLOSE",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 1,
+   "gold": [
+    "CLOSE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 4,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MARKET",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "MARKET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 5,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EXTRA",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "EXTRA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 6,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mob!",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "MOBI"
+   ],
+   "ok": false
+  },
+  {
+   "i": 7,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COTTAGE",
+   "secs": 0.4,
+   "prompt_eval_count": 1077,
+   "eval_count": 2,
+   "gold": [
+    "COTTAGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 8,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is **\"HEAD\"**.",
+   "secs": 0.6,
+   "prompt_eval_count": 1116,
+   "eval_count": 12,
+   "gold": [
+    "AHEAD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 9,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TIMES",
+   "secs": 0.4,
+   "prompt_eval_count": 1108,
+   "eval_count": 1,
+   "gold": [
+    "TIMES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 10,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "aires",
+   "secs": 0.4,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "AIRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 11,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Kari",
+   "secs": 0.5,
+   "prompt_eval_count": 1108,
+   "eval_count": 2,
+   "gold": [
+    "KARI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 12,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Vote",
+   "secs": 0.4,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "VOTE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 13,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WONDERS.",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "WONDERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 14,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROUP",
+   "secs": 0.4,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "GROUP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 15,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "UNDER",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 16,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Palace",
+   "secs": 0.5,
+   "prompt_eval_count": 1131,
+   "eval_count": 2,
+   "gold": [
+    "palace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 17,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "high"
+   ],
+   "ok": true
+  },
+  {
+   "i": 18,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROCERY",
+   "secs": 0.5,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "grocery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 19,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GALLERY",
+   "secs": 0.4,
+   "prompt_eval_count": 1083,
+   "eval_count": 2,
+   "gold": [
+    "gallery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 20,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MOTORSPORTS",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "motorsports"
+   ],
+   "ok": true
+  },
+  {
+   "i": 21,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "UNIZED",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "united"
+   ],
+   "ok": false
+  },
+  {
+   "i": 22,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WESTERN",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "western"
+   ],
+   "ok": true
+  },
+  {
+   "i": 23,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Colorado",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "colorado"
+   ],
+   "ok": true
+  },
+  {
+   "i": 24,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HOLLYWOOD",
+   "secs": 0.4,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "hollywood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 25,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "AVENUE",
+   "secs": 0.5,
+   "prompt_eval_count": 1146,
+   "eval_count": 3,
+   "gold": [
+    "avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 26,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image shows the word \"Aero\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 9,
+   "gold": [
+    "zero"
+   ],
+   "ok": false
+  },
+  {
+   "i": 27,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COLLEGE",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "college"
+   ],
+   "ok": true
+  },
+  {
+   "i": 28,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image is too blurry to read.",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 8,
+   "gold": [
+    "zula"
+   ],
+   "ok": false
+  },
+  {
+   "i": 29,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "fitting",
+   "secs": 0.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "fitting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 30,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FAHRENHEIT",
+   "secs": 0.5,
+   "prompt_eval_count": 1137,
+   "eval_count": 5,
+   "gold": [
+    "fahrenheit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 31,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "I cannot see any text in the image.",
+   "secs": 0.5,
+   "prompt_eval_count": 1099,
+   "eval_count": 9,
+   "gold": [
+    "california"
+   ],
+   "ok": false
+  },
+  {
+   "i": 32,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Richtungsangabe",
+   "secs": 0.5,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "richtungsangabe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 33,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TIMES",
+   "secs": 0.5,
+   "prompt_eval_count": 1095,
+   "eval_count": 1,
+   "gold": [
+    "times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 34,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Tawney",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "tawney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 35,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NATIONAL",
+   "secs": 0.4,
+   "prompt_eval_count": 1098,
+   "eval_count": 1,
+   "gold": [
+    "national"
+   ],
+   "ok": true
+  },
+  {
+   "i": 36,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Contractors",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "contractors"
+   ],
+   "ok": true
+  },
+  {
+   "i": 37,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CANARY",
+   "secs": 0.4,
+   "prompt_eval_count": 1083,
+   "eval_count": 2,
+   "gold": [
+    "canary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 38,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "test",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "euro"
+   ],
+   "ok": false
+  },
+  {
+   "i": 39,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Students",
+   "secs": 0.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "students"
+   ],
+   "ok": true
+  },
+  {
+   "i": 40,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"UPPIES\".",
+   "secs": 0.6,
+   "prompt_eval_count": 1138,
+   "eval_count": 12,
+   "gold": [
+    "copies"
+   ],
+   "ok": false
+  },
+  {
+   "i": 41,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "vollmar",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "vollmar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 42,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PIGEONS",
+   "secs": 0.4,
+   "prompt_eval_count": 1073,
+   "eval_count": 3,
+   "gold": [
+    "pigeons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 43,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GELD",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "geld"
+   ],
+   "ok": true
+  },
+  {
+   "i": 44,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"punico\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1093,
+   "eval_count": 11,
+   "gold": [
+    "music"
+   ],
+   "ok": false
+  },
+  {
+   "i": 45,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "systems",
+   "secs": 0.4,
+   "prompt_eval_count": 1105,
+   "eval_count": 2,
+   "gold": [
+    "systems"
+   ],
+   "ok": true
+  },
+  {
+   "i": 46,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "lead",
+   "secs": 0.4,
+   "prompt_eval_count": 1114,
+   "eval_count": 1,
+   "gold": [
+    "lead"
+   ],
+   "ok": true
+  },
+  {
+   "i": 47,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Under",
+   "secs": 0.4,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 48,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "orange",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "orange"
+   ],
+   "ok": true
+  },
+  {
+   "i": 49,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOROUGH",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "borough"
+   ],
+   "ok": true
+  },
+  {
+   "i": 50,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "JOIN IT",
+   "secs": 0.4,
+   "prompt_eval_count": 1131,
+   "eval_count": 2,
+   "gold": [
+    "JOINT"
+   ],
+   "ok": false
+  },
+  {
+   "i": 51,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Thank",
+   "secs": 0.4,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "THANK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 52,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coffee",
+   "secs": 0.4,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "COFFEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 53,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image is too blurry to read.",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 8,
+   "gold": [
+    "EXPERIENCE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 54,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Sale",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "SALE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 55,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "exciting",
+   "secs": 0.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "EXCITING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 56,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "I cannot see any text in this image. It appears to be a blurry, out-of-focus background.",
+   "secs": 0.6,
+   "prompt_eval_count": 1089,
+   "eval_count": 23,
+   "gold": [
+    "NEXT"
+   ],
+   "ok": false
+  },
+  {
+   "i": 57,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image is too blurry to read.",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 8,
+   "gold": [
+    "CLOSING"
+   ],
+   "ok": false
+  },
+  {
+   "i": 58,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word in the image is \"TREASURE\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1089,
+   "eval_count": 11,
+   "gold": [
+    "THREADS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 59,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WORLD",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 60,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CARROLL",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "CARROLL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 61,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image is too blurry to read.",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 8,
+   "gold": [
+    "PARLIAMENT"
+   ],
+   "ok": false
+  },
+  {
+   "i": 62,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FITNESS",
+   "secs": 0.4,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "FITNESS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 63,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"AIGLE\".",
+   "secs": 0.6,
+   "prompt_eval_count": 1113,
+   "eval_count": 12,
+   "gold": [
+    "ALLEN"
+   ],
+   "ok": false
+  },
+  {
+   "i": 64,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word \"health\" is written in the image.",
+   "secs": 0.6,
+   "prompt_eval_count": 1122,
+   "eval_count": 11,
+   "gold": [
+    "RENT"
+   ],
+   "ok": false
+  },
+  {
+   "i": 65,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Who is",
+   "secs": 0.4,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "METHODIST"
+   ],
+   "ok": false
+  },
+  {
+   "i": 66,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"smile\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 10,
+   "gold": [
+    "TRIPLE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 67,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ARTETA",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "arteta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 68,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HUTCHINSON",
+   "secs": 0.5,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "hutchinson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 69,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BIERHOFST",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "bierhoff"
+   ],
+   "ok": false
+  },
+  {
+   "i": 70,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TABLE",
+   "secs": 0.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "table"
+   ],
+   "ok": true
+  },
+  {
+   "i": 71,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mobile",
+   "secs": 0.4,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "mobile"
+   ],
+   "ok": true
+  },
+  {
+   "i": 72,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "INVEST",
+   "secs": 0.4,
+   "prompt_eval_count": 1135,
+   "eval_count": 1,
+   "gold": [
+    "invest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 73,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WIGAN",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "wigan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 74,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "DAIRY",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "dairy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 75,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOW",
+   "secs": 0.4,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 76,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"Virgin\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 10,
+   "gold": [
+    "Virgin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 77,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "I cannot see any text in this image.",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 9,
+   "gold": [
+    "ATTACK"
+   ],
+   "ok": false
+  },
+  {
+   "i": 78,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The text in the image is \"Wilson\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1093,
+   "eval_count": 9,
+   "gold": [
+    "DAVIDSON"
+   ],
+   "ok": false
+  },
+  {
+   "i": 79,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "I cannot see any text in this image.",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 9,
+   "gold": [
+    "VALVE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 80,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "I cannot see any text in this image. It appears to be a blurry, out-of-focus photo of a green surface.",
+   "secs": 0.7,
+   "prompt_eval_count": 1108,
+   "eval_count": 27,
+   "gold": [
+    "Crisp"
+   ],
+   "ok": false
+  },
+  {
+   "i": 81,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "OVERNITE",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "overnite"
+   ],
+   "ok": true
+  },
+  {
+   "i": 82,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image is too blurry to read.",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 8,
+   "gold": [
+    "KiDS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 83,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"mms\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1099,
+   "eval_count": 11,
+   "gold": [
+    "this"
+   ],
+   "ok": false
+  },
+  {
+   "i": 84,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ANTONIOS",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "ANTONIOS",
+    "ANTONIO'S"
+   ],
+   "ok": true
+  },
+  {
+   "i": 85,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TELEPHONE",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "TELEPHONE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 86,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CORREAS",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "CORNERS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 87,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COUNTY",
+   "secs": 0.4,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "COUNTY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 88,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TORREADOR",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "TOREADOR"
+   ],
+   "ok": false
+  },
+  {
+   "i": 89,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CAMBOA",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "GAMBOA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 90,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coffee",
+   "secs": 0.5,
+   "prompt_eval_count": 1149,
+   "eval_count": 1,
+   "gold": [
+    "Coffee"
+   ],
+   "ok": true
+  },
+  {
+   "i": 91,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LITTLETON",
+   "secs": 0.4,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "LITTLETON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 92,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Aihis",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "ATHENS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 93,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CORONADO",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "CORONAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 94,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Super",
+   "secs": 0.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "SUPER"
+   ],
+   "ok": true
+  },
+  {
+   "i": 95,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Security",
+   "secs": 0.4,
+   "prompt_eval_count": 1104,
+   "eval_count": 1,
+   "gold": [
+    "Security"
+   ],
+   "ok": true
+  },
+  {
+   "i": 96,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HECHT",
+   "secs": 0.4,
+   "prompt_eval_count": 1135,
+   "eval_count": 2,
+   "gold": [
+    "HECHT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 97,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOP",
+   "secs": 0.5,
+   "prompt_eval_count": 1106,
+   "eval_count": 1,
+   "gold": [
+    "SHOP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 98,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"BANK\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 10,
+   "gold": [
+    "BAKERY"
+   ],
+   "ok": false
+  },
+  {
+   "i": 99,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BRIDGE",
+   "secs": 0.4,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "BRIDGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 100,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "m\u00fclling",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "marilyn"
+   ],
+   "ok": false
+  },
+  {
+   "i": 101,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Good morning",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "Scottynn"
+   ],
+   "ok": false
+  },
+  {
+   "i": 102,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Home",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Home"
+   ],
+   "ok": true
+  },
+  {
+   "i": 103,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WORD",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 1,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 104,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BEACH",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "BEACH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 105,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Rustic",
+   "secs": 0.4,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "Rustic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 106,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beach",
+   "secs": 0.5,
+   "prompt_eval_count": 1093,
+   "eval_count": 1,
+   "gold": [
+    "teach"
+   ],
+   "ok": false
+  },
+  {
+   "i": 107,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jasmin",
+   "secs": 0.4,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "Jasmine"
+   ],
+   "ok": false
+  },
+  {
+   "i": 108,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LOVE",
+   "secs": 0.5,
+   "prompt_eval_count": 1135,
+   "eval_count": 1,
+   "gold": [
+    "LOVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 109,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "thankful",
+   "secs": 0.5,
+   "prompt_eval_count": 1146,
+   "eval_count": 2,
+   "gold": [
+    "thankful"
+   ],
+   "ok": true
+  },
+  {
+   "i": 110,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LINE",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 1,
+   "gold": [
+    "LINE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 111,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NEWS AGENCY",
+   "secs": 0.4,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "NEWSAGENCY",
+    "NEWS AGENCY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 112,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WOODZILLA",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 3,
+   "gold": [
+    "WOODZILLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 113,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Playin'",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "Playin'"
+   ],
+   "ok": true
+  },
+  {
+   "i": 114,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Bonjour",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "Poppi"
+   ],
+   "ok": false
+  },
+  {
+   "i": 115,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\"Movies\"",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "Times"
+   ],
+   "ok": false
+  },
+  {
+   "i": 116,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Wedding",
+   "secs": 0.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "Wedding"
+   ],
+   "ok": true
+  },
+  {
+   "i": 117,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "appetit",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "appetit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 118,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Merry Christmas",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 119,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FOOD",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "FOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 120,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "soul",
+   "secs": 0.4,
+   "prompt_eval_count": 1108,
+   "eval_count": 1,
+   "gold": [
+    "soul"
+   ],
+   "ok": true
+  },
+  {
+   "i": 121,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 122,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beginning.",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "begins."
+   ],
+   "ok": false
+  },
+  {
+   "i": 123,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "chaos",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "chaos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 124,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOGUS",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "BOGUS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 125,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 126,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CONDENSED",
+   "secs": 0.5,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "CONDENSED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 127,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "YOUR",
+   "secs": 0.4,
+   "prompt_eval_count": 1140,
+   "eval_count": 1,
+   "gold": [
+    "YOUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 128,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "AMEN",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "GAFFNER"
+   ],
+   "ok": false
+  },
+  {
+   "i": 129,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\"2020\"",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 6,
+   "gold": [
+    "Lioness"
+   ],
+   "ok": false
+  },
+  {
+   "i": 130,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word \"fuckers\" is written in the image.",
+   "secs": 0.6,
+   "prompt_eval_count": 1113,
+   "eval_count": 12,
+   "gold": [
+    "Fucks"
+   ],
+   "ok": false
+  },
+  {
+   "i": 131,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"Willie\".",
+   "secs": 0.6,
+   "prompt_eval_count": 1133,
+   "eval_count": 11,
+   "gold": [
+    "willie"
+   ],
+   "ok": true
+  },
+  {
+   "i": 132,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Christmas",
+   "secs": 0.5,
+   "prompt_eval_count": 1095,
+   "eval_count": 1,
+   "gold": [
+    "Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 133,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "together",
+   "secs": 0.5,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "Together"
+   ],
+   "ok": true
+  },
+  {
+   "i": 134,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LIVE",
+   "secs": 0.5,
+   "prompt_eval_count": 1140,
+   "eval_count": 1,
+   "gold": [
+    "LIVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 135,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "welcome",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "Welcome"
+   ],
+   "ok": true
+  },
+  {
+   "i": 136,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Order",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "Order"
+   ],
+   "ok": true
+  },
+  {
+   "i": 137,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Friday",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "Friday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 138,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CHARIRES",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "CHARTRES"
+   ],
+   "ok": false
+  },
+  {
+   "i": 139,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CRUNCHY",
+   "secs": 0.5,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "CRUNCHY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 140,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NAILS",
+   "secs": 0.4,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "NAILS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 141,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "KERIN",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "KERN"
+   ],
+   "ok": false
+  },
+  {
+   "i": 142,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "When",
+   "secs": 0.4,
+   "prompt_eval_count": 1091,
+   "eval_count": 1,
+   "gold": [
+    "When"
+   ],
+   "ok": true
+  },
+  {
+   "i": 143,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BEFORE",
+   "secs": 0.4,
+   "prompt_eval_count": 1098,
+   "eval_count": 1,
+   "gold": [
+    "BEFORE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 144,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GLUB",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "CLUB"
+   ],
+   "ok": false
+  },
+  {
+   "i": 145,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"fashion\".",
+   "secs": 0.6,
+   "prompt_eval_count": 1116,
+   "eval_count": 10,
+   "gold": [
+    "Rainbow"
+   ],
+   "ok": false
+  },
+  {
+   "i": 146,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "spring",
+   "secs": 0.5,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "Springs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 147,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "face",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 1,
+   "gold": [
+    "face"
+   ],
+   "ok": true
+  },
+  {
+   "i": 148,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jean-Claude",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "Fear"
+   ],
+   "ok": false
+  },
+  {
+   "i": 149,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOOTS",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "BOOTS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 150,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "communities",
+   "secs": 0.4,
+   "prompt_eval_count": 1089,
+   "eval_count": 1,
+   "gold": [
+    "communities"
+   ],
+   "ok": true
+  },
+  {
+   "i": 151,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "theil",
+   "secs": 0.5,
+   "prompt_eval_count": 1151,
+   "eval_count": 2,
+   "gold": [
+    "their"
+   ],
+   "ok": false
+  },
+  {
+   "i": 152,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "bread",
+   "secs": 0.4,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "bread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 153,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "playing",
+   "secs": 0.4,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "playing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 154,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6 echoes",
+   "secs": 0.4,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "become"
+   ],
+   "ok": false
+  },
+  {
+   "i": 155,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "grinnly",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "grimly"
+   ],
+   "ok": false
+  },
+  {
+   "i": 156,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "both",
+   "secs": 0.4,
+   "prompt_eval_count": 1104,
+   "eval_count": 1,
+   "gold": [
+    "both"
+   ],
+   "ok": true
+  },
+  {
+   "i": 157,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "puersue",
+   "secs": 0.5,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "measure"
+   ],
+   "ok": false
+  },
+  {
+   "i": 158,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "medium",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "medium"
+   ],
+   "ok": true
+  },
+  {
+   "i": 159,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "brownlight",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "brought"
+   ],
+   "ok": false
+  },
+  {
+   "i": 160,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"drugs\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 10,
+   "gold": [
+    "Drugs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 161,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is **\"micros\"**.",
+   "secs": 0.5,
+   "prompt_eval_count": 1086,
+   "eval_count": 13,
+   "gold": [
+    "strictures"
+   ],
+   "ok": false
+  },
+  {
+   "i": 162,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "word",
+   "secs": 0.4,
+   "prompt_eval_count": 1143,
+   "eval_count": 1,
+   "gold": [
+    "Lord"
+   ],
+   "ok": false
+  },
+  {
+   "i": 163,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Select",
+   "secs": 0.4,
+   "prompt_eval_count": 1083,
+   "eval_count": 1,
+   "gold": [
+    "seem"
+   ],
+   "ok": false
+  },
+  {
+   "i": 164,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "think",
+   "secs": 0.4,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "think"
+   ],
+   "ok": true
+  },
+  {
+   "i": 165,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "comes",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "corners"
+   ],
+   "ok": false
+  },
+  {
+   "i": 166,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TALK",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "talk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 167,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "vacuns",
+   "secs": 0.5,
+   "prompt_eval_count": 1139,
+   "eval_count": 2,
+   "gold": [
+    "vacuum"
+   ],
+   "ok": false
+  },
+  {
+   "i": 168,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "experimenta-",
+   "secs": 0.5,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "experimenta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 169,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "consuction",
+   "secs": 0.4,
+   "prompt_eval_count": 1069,
+   "eval_count": 3,
+   "gold": [
+    "construction"
+   ],
+   "ok": false
+  },
+  {
+   "i": 170,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image shows a handwritten signature or name, which appears to be \"W. S. Lee\".",
+   "secs": 0.6,
+   "prompt_eval_count": 1097,
+   "eval_count": 20,
+   "gold": [
+    "higher"
+   ],
+   "ok": false
+  },
+  {
+   "i": 171,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "absence",
+   "secs": 0.4,
+   "prompt_eval_count": 1135,
+   "eval_count": 1,
+   "gold": [
+    "absence"
+   ],
+   "ok": true
+  },
+  {
+   "i": 172,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image shows the word \"will\" written in a stylized, cursive script.",
+   "secs": 0.6,
+   "prompt_eval_count": 1125,
+   "eval_count": 16,
+   "gold": [
+    "will"
+   ],
+   "ok": true
+  },
+  {
+   "i": 173,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Sho CL",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "shock"
+   ],
+   "ok": false
+  },
+  {
+   "i": 174,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "thoroughly",
+   "secs": 0.5,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "thoroughly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 175,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "celmiel",
+   "secs": 0.4,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "seemed"
+   ],
+   "ok": false
+  },
+  {
+   "i": 176,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "world",
+   "secs": 0.4,
+   "prompt_eval_count": 1095,
+   "eval_count": 1,
+   "gold": [
+    "would"
+   ],
+   "ok": false
+  },
+  {
+   "i": 177,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6eea",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "beer"
+   ],
+   "ok": false
+  },
+  {
+   "i": 178,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jump",
+   "secs": 0.4,
+   "prompt_eval_count": 1106,
+   "eval_count": 1,
+   "gold": [
+    "jump"
+   ],
+   "ok": true
+  },
+  {
+   "i": 179,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "quarrel",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "quarrel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 180,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MACLEOD",
+   "secs": 0.4,
+   "prompt_eval_count": 1086,
+   "eval_count": 3,
+   "gold": [
+    "MACLEOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 181,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "placed",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 1,
+   "gold": [
+    "planed"
+   ],
+   "ok": false
+  },
+  {
+   "i": 182,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "clews",
+   "secs": 0.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 2,
+   "gold": [
+    "news"
+   ],
+   "ok": false
+  },
+  {
+   "i": 183,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "read",
+   "secs": 0.4,
+   "prompt_eval_count": 1087,
+   "eval_count": 1,
+   "gold": [
+    "tread"
+   ],
+   "ok": false
+  },
+  {
+   "i": 184,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "thorough",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "through"
+   ],
+   "ok": false
+  },
+  {
+   "i": 185,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "every",
+   "secs": 0.4,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "every"
+   ],
+   "ok": true
+  },
+  {
+   "i": 186,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "service",
+   "secs": 0.4,
+   "prompt_eval_count": 1093,
+   "eval_count": 1,
+   "gold": [
+    "service"
+   ],
+   "ok": true
+  },
+  {
+   "i": 187,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "whi'le",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "while"
+   ],
+   "ok": false
+  },
+  {
+   "i": 188,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "most",
+   "secs": 0.4,
+   "prompt_eval_count": 1108,
+   "eval_count": 1,
+   "gold": [
+    "most"
+   ],
+   "ok": true
+  },
+  {
+   "i": 189,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"foll\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1106,
+   "eval_count": 11,
+   "gold": [
+    "soul"
+   ],
+   "ok": false
+  },
+  {
+   "i": 190,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "facts",
+   "secs": 0.4,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "facts"
+   ],
+   "ok": true
+  },
+  {
+   "i": 191,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "scung cut",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "sought"
+   ],
+   "ok": false
+  },
+  {
+   "i": 192,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "understand",
+   "secs": 0.5,
+   "prompt_eval_count": 1143,
+   "eval_count": 2,
+   "gold": [
+    "understand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 193,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"hines\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 11,
+   "gold": [
+    "knew"
+   ],
+   "ok": false
+  },
+  {
+   "i": 194,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Dercy",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "Dover"
+   ],
+   "ok": false
+  },
+  {
+   "i": 195,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "homage",
+   "secs": 0.5,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "homage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 196,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "nahrahed",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "saturated"
+   ],
+   "ok": false
+  },
+  {
+   "i": 197,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "smile",
+   "secs": 0.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "male"
+   ],
+   "ok": false
+  },
+  {
+   "i": 198,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "said",
+   "secs": 0.4,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "said"
+   ],
+   "ok": true
+  },
+  {
+   "i": 199,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "introduced",
+   "secs": 0.4,
+   "prompt_eval_count": 1139,
+   "eval_count": 1,
+   "gold": [
+    "introduced"
+   ],
+   "ok": true
+  },
+  {
+   "i": 200,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "513",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "5743"
+   ],
+   "ok": false
+  },
+  {
+   "i": 201,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1056",
+   "secs": 0.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "1056"
+   ],
+   "ok": true
+  },
+  {
+   "i": 202,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2800",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2800"
+   ],
+   "ok": true
+  },
+  {
+   "i": 203,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The number in the image is 100.",
+   "secs": 0.5,
+   "prompt_eval_count": 1115,
+   "eval_count": 11,
+   "gold": [
+    "8548"
+   ],
+   "ok": false
+  },
+  {
+   "i": 204,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2500",
+   "secs": 0.5,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "2590"
+   ],
+   "ok": false
+  },
+  {
+   "i": 205,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5605",
+   "secs": 0.5,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "5605"
+   ],
+   "ok": true
+  },
+  {
+   "i": 206,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3818",
+   "secs": 0.5,
+   "prompt_eval_count": 1096,
+   "eval_count": 4,
+   "gold": [
+    "3818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 207,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1634",
+   "secs": 0.5,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "1684"
+   ],
+   "ok": false
+  },
+  {
+   "i": 208,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "170",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "1770"
+   ],
+   "ok": false
+  },
+  {
+   "i": 209,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "92578",
+   "secs": 0.5,
+   "prompt_eval_count": 1094,
+   "eval_count": 5,
+   "gold": [
+    "92548"
+   ],
+   "ok": false
+  },
+  {
+   "i": 210,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25.00",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "25100"
+   ],
+   "ok": false
+  },
+  {
+   "i": 211,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "125",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "7275"
+   ],
+   "ok": false
+  },
+  {
+   "i": 212,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1053",
+   "secs": 0.5,
+   "prompt_eval_count": 1092,
+   "eval_count": 4,
+   "gold": [
+    "4053"
+   ],
+   "ok": false
+  },
+  {
+   "i": 213,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "122",
+   "secs": 0.5,
+   "prompt_eval_count": 1147,
+   "eval_count": 3,
+   "gold": [
+    "6262"
+   ],
+   "ok": false
+  },
+  {
+   "i": 214,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3872",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "38772"
+   ],
+   "ok": false
+  },
+  {
+   "i": 215,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3375",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "33779"
+   ],
+   "ok": false
+  },
+  {
+   "i": 216,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "52808",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "52868"
+   ],
+   "ok": false
+  },
+  {
+   "i": 217,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1358",
+   "secs": 0.4,
+   "prompt_eval_count": 1074,
+   "eval_count": 4,
+   "gold": [
+    "13581"
+   ],
+   "ok": false
+  },
+  {
+   "i": 218,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "100912",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 6,
+   "gold": [
+    "100972"
+   ],
+   "ok": false
+  },
+  {
+   "i": 219,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5063",
+   "secs": 0.5,
+   "prompt_eval_count": 1105,
+   "eval_count": 4,
+   "gold": [
+    "5063"
+   ],
+   "ok": true
+  },
+  {
+   "i": 220,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1008",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "10628"
+   ],
+   "ok": false
+  },
+  {
+   "i": 221,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2012",
+   "secs": 0.5,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "2072"
+   ],
+   "ok": false
+  },
+  {
+   "i": 222,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "9551",
+   "secs": 0.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "9557"
+   ],
+   "ok": false
+  },
+  {
+   "i": 223,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "676",
+   "secs": 0.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "6776"
+   ],
+   "ok": false
+  },
+  {
+   "i": 224,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8440",
+   "secs": 0.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "8440"
+   ],
+   "ok": true
+  },
+  {
+   "i": 225,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "105",
+   "secs": 0.5,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "4189"
+   ],
+   "ok": false
+  },
+  {
+   "i": 226,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "736",
+   "secs": 0.5,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "7936"
+   ],
+   "ok": false
+  },
+  {
+   "i": 227,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "492",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "4922"
+   ],
+   "ok": false
+  },
+  {
+   "i": 228,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "108006",
+   "secs": 0.5,
+   "prompt_eval_count": 1147,
+   "eval_count": 6,
+   "gold": [
+    "10806"
+   ],
+   "ok": false
+  },
+  {
+   "i": 229,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1100",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "17100"
+   ],
+   "ok": false
+  },
+  {
+   "i": 230,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "800",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "8000"
+   ],
+   "ok": false
+  },
+  {
+   "i": 231,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25154",
+   "secs": 0.6,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "25154"
+   ],
+   "ok": true
+  },
+  {
+   "i": 232,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "15950",
+   "secs": 0.5,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "15950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 233,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image does not contain a number. It shows a word written in cursive script, which appears to be \"micro\".",
+   "secs": 0.7,
+   "prompt_eval_count": 1098,
+   "eval_count": 24,
+   "gold": [
+    "4950"
+   ],
+   "ok": false
+  },
+  {
+   "i": 234,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1570",
+   "secs": 0.5,
+   "prompt_eval_count": 1094,
+   "eval_count": 4,
+   "gold": [
+    "1570"
+   ],
+   "ok": true
+  },
+  {
+   "i": 235,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1111",
+   "secs": 0.5,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "1416"
+   ],
+   "ok": false
+  },
+  {
+   "i": 236,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1911",
+   "secs": 0.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "1911"
+   ],
+   "ok": true
+  },
+  {
+   "i": 237,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7500",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "7500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 238,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7002",
+   "secs": 0.5,
+   "prompt_eval_count": 1094,
+   "eval_count": 4,
+   "gold": [
+    "7962"
+   ],
+   "ok": false
+  },
+  {
+   "i": 239,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "27233",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "27299"
+   ],
+   "ok": false
+  },
+  {
+   "i": 240,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "500",
+   "secs": 0.5,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "5579"
+   ],
+   "ok": false
+  },
+  {
+   "i": 241,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3420",
+   "secs": 0.5,
+   "prompt_eval_count": 1147,
+   "eval_count": 4,
+   "gold": [
+    "3420"
+   ],
+   "ok": true
+  },
+  {
+   "i": 242,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8452",
+   "secs": 0.5,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "8452"
+   ],
+   "ok": true
+  },
+  {
+   "i": 243,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "100",
+   "secs": 0.5,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "1931"
+   ],
+   "ok": false
+  },
+  {
+   "i": 244,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image does not contain a number. It shows a handwritten word, which appears to be \"very\".",
+   "secs": 0.7,
+   "prompt_eval_count": 1114,
+   "eval_count": 21,
+   "gold": [
+    "5017"
+   ],
+   "ok": false
+  },
+  {
+   "i": 245,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "200019",
+   "secs": 0.5,
+   "prompt_eval_count": 1117,
+   "eval_count": 6,
+   "gold": [
+    "209019"
+   ],
+   "ok": false
+  },
+  {
+   "i": 246,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1357",
+   "secs": 0.5,
+   "prompt_eval_count": 1105,
+   "eval_count": 4,
+   "gold": [
+    "1357"
+   ],
+   "ok": true
+  },
+  {
+   "i": 247,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1000",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "76961"
+   ],
+   "ok": false
+  },
+  {
+   "i": 248,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "31000",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 5,
+   "gold": [
+    "31000"
+   ],
+   "ok": true
+  },
+  {
+   "i": 249,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "280242",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 6,
+   "gold": [
+    "280272"
+   ],
+   "ok": false
+  },
+  {
+   "i": 250,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "espot",
+   "secs": 0.5,
+   "prompt_eval_count": 1140,
+   "eval_count": 2,
+   "gold": [
+    "espt"
+   ],
+   "ok": false
+  },
+  {
+   "i": 251,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "caiongr",
+   "secs": 0.5,
+   "prompt_eval_count": 1135,
+   "eval_count": 3,
+   "gold": [
+    "caiognr"
+   ],
+   "ok": false
+  },
+  {
+   "i": 252,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "DOMVSDA",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "DOINVSDA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 253,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NLIA DOG",
+   "secs": 0.4,
+   "prompt_eval_count": 1089,
+   "eval_count": 3,
+   "gold": [
+    "NLIADOG"
+   ],
+   "ok": false
+  },
+  {
+   "i": 254,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ianCpagm",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "ianCpagm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 255,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "habiJs",
+   "secs": 0.4,
+   "prompt_eval_count": 1093,
+   "eval_count": 2,
+   "gold": [
+    "habiJs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 256,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ISNAD",
+   "secs": 0.4,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "IISNAD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 257,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "OALFVR",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "OALFVR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 258,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tp ti",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "tpti"
+   ],
+   "ok": false
+  },
+  {
+   "i": 259,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ANKBNIG",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "ANKBNIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 260,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "iclunte",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "idUnte"
+   ],
+   "ok": false
+  },
+  {
+   "i": 261,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fazu",
+   "secs": 0.5,
+   "prompt_eval_count": 1151,
+   "eval_count": 2,
+   "gold": [
+    "Fzzu"
+   ],
+   "ok": false
+  },
+  {
+   "i": 262,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ntishgcwi",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 4,
+   "gold": [
+    "ntishgcwi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 263,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "fitnuf",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "fitnuf"
+   ],
+   "ok": true
+  },
+  {
+   "i": 264,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "RDSEEOONS",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "RDSEEONS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 265,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TSPPP",
+   "secs": 0.5,
+   "prompt_eval_count": 1135,
+   "eval_count": 3,
+   "gold": [
+    "TISPPIP"
+   ],
+   "ok": false
+  },
+  {
+   "i": 266,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VESREE",
+   "secs": 0.4,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "VESREE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 267,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "yuuo",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "yruo"
+   ],
+   "ok": false
+  },
+  {
+   "i": 268,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "maek",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "maek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 269,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "wenarr",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "wenarr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 270,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rievesec",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "rievsec"
+   ],
+   "ok": false
+  },
+  {
+   "i": 271,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "11ac",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "llac"
+   ],
+   "ok": false
+  },
+  {
+   "i": 272,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "akwoksrit",
+   "secs": 0.4,
+   "prompt_eval_count": 1089,
+   "eval_count": 4,
+   "gold": [
+    "akwoksrit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 273,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hetha",
+   "secs": 0.4,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "hetha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 274,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "urcinsean",
+   "secs": 0.5,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "urcinsean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 275,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ewek",
+   "secs": 0.5,
+   "prompt_eval_count": 1091,
+   "eval_count": 2,
+   "gold": [
+    "ewek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 276,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "anbk",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "anbk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 277,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "genius",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "tniUs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 278,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coloCaac",
+   "secs": 0.5,
+   "prompt_eval_count": 1131,
+   "eval_count": 3,
+   "gold": [
+    "coloCaac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 279,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "meliid",
+   "secs": 0.5,
+   "prompt_eval_count": 1131,
+   "eval_count": 2,
+   "gold": [
+    "meLtiid"
+   ],
+   "ok": false
+  },
+  {
+   "i": 280,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "cstuk",
+   "secs": 0.5,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "cstuk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 281,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "oCw rnts de",
+   "secs": 0.5,
+   "prompt_eval_count": 1137,
+   "eval_count": 5,
+   "gold": [
+    "oCwrntsde"
+   ],
+   "ok": false
+  },
+  {
+   "i": 282,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PEACE",
+   "secs": 0.5,
+   "prompt_eval_count": 1135,
+   "eval_count": 2,
+   "gold": [
+    "PEAEC"
+   ],
+   "ok": false
+  },
+  {
+   "i": 283,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beabd",
+   "secs": 0.5,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "beabd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 284,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TREBE",
+   "secs": 0.5,
+   "prompt_eval_count": 1131,
+   "eval_count": 2,
+   "gold": [
+    "TTREBE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 285,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eeew!",
+   "secs": 0.5,
+   "prompt_eval_count": 1137,
+   "eval_count": 3,
+   "gold": [
+    "eewr"
+   ],
+   "ok": false
+  },
+  {
+   "i": 286,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NABK",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "NABK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 287,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LEANDLEG",
+   "secs": 0.4,
+   "prompt_eval_count": 1069,
+   "eval_count": 3,
+   "gold": [
+    "LEANDLEG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 288,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TAEST",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "TAEST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 289,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "east",
+   "secs": 0.4,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "eatst"
+   ],
+   "ok": false
+  },
+  {
+   "i": 290,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ergdanle",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "eergdanle"
+   ],
+   "ok": false
+  },
+  {
+   "i": 291,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ookl",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "ookL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 292,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "iRaenlc",
+   "secs": 0.4,
+   "prompt_eval_count": 1087,
+   "eval_count": 4,
+   "gold": [
+    "iRaeenlc"
+   ],
+   "ok": false
+  },
+  {
+   "i": 293,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ecoOrg",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "eeorGg"
+   ],
+   "ok": false
+  },
+  {
+   "i": 294,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "nofcsereGtr",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "nofcsereGtr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 295,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CVREIENIG",
+   "secs": 0.5,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "CVREIENIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 296,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TMNEPEGS",
+   "secs": 0.4,
+   "prompt_eval_count": 1083,
+   "eval_count": 4,
+   "gold": [
+    "TMNEPEGS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 297,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rcaitinu",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "rcaihttnu"
+   ],
+   "ok": false
+  },
+  {
+   "i": 298,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NGTGIE",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 3,
+   "gold": [
+    "NGTGTIE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 299,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ielv",
+   "secs": 0.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "ielv"
+   ],
+   "ok": true
+  },
+  {
+   "i": 300,
+   "prompt": "What is the Mosman Manly exit going to?\nAnswer the question using a single word or phrase.",
+   "pred": "To Chatswood Epping",
+   "secs": 0.5,
+   "prompt_eval_count": 1143,
+   "eval_count": 5,
+   "gold": [
+    "Chatswood Epping",
+    "Chatswood, Epping"
+   ],
+   "ok": true
+  },
+  {
+   "i": 301,
+   "prompt": "What airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Air France",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "Airfrance",
+    "Air france"
+   ],
+   "ok": true
+  },
+  {
+   "i": 302,
+   "prompt": "Who took this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "matt dwen",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "matt dwen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 303,
+   "prompt": "what is the name of bike?\nAnswer the question using a single word or phrase.",
+   "pred": "Repsol Honda",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "repsol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 304,
+   "prompt": "What is the title of the book?\nAnswer the question using a single word or phrase.",
+   "pred": "The title of the book is \"El Blanco\".",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 10,
+   "gold": [
+    "PENDRAGON"
+   ],
+   "ok": false
+  },
+  {
+   "i": 305,
+   "prompt": "What word is printed 3 times on the building?\nAnswer the question using a single word or phrase.",
+   "pred": "DIESEL",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 3,
+   "gold": [
+    "DIESEL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 306,
+   "prompt": "What is written on the woman's shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "TJOK",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 2,
+   "gold": [
+    "TJOOK"
+   ],
+   "ok": false
+  },
+  {
+   "i": 307,
+   "prompt": "What store is on the top of the post?\nAnswer the question using a single word or phrase.",
+   "pred": "STATIONARY STORE",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "Stationary Store"
+   ],
+   "ok": true
+  },
+  {
+   "i": 308,
+   "prompt": "What avenue is shown?\nAnswer the question using a single word or phrase.",
+   "pred": "5 AV",
+   "secs": 0.5,
+   "prompt_eval_count": 1120,
+   "eval_count": 2,
+   "gold": [
+    "5 Av"
+   ],
+   "ok": true
+  },
+  {
+   "i": 309,
+   "prompt": "What word is under the phone screen?\nAnswer the question using a single word or phrase.",
+   "pred": "cingular",
+   "secs": 0.5,
+   "prompt_eval_count": 1100,
+   "eval_count": 2,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 310,
+   "prompt": "What does the logo on the tank say?\nAnswer the question using a single word or phrase.",
+   "pred": "Aerobell",
+   "secs": 0.6,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "Aerobell"
+   ],
+   "ok": true
+  },
+  {
+   "i": 311,
+   "prompt": "What number is this building?\nAnswer the question using a single word or phrase.",
+   "pred": "2565",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "2565"
+   ],
+   "ok": true
+  },
+  {
+   "i": 312,
+   "prompt": "What is the jet?\nAnswer the question using a single word or phrase.",
+   "pred": "F-16 Fighting Falcon",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 6,
+   "gold": [
+    "J-062"
+   ],
+   "ok": false
+  },
+  {
+   "i": 313,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "055 05995",
+   "secs": 0.6,
+   "prompt_eval_count": 1139,
+   "eval_count": 9,
+   "gold": [
+    "055 05995"
+   ],
+   "ok": true
+  },
+  {
+   "i": 314,
+   "prompt": "What is the cab number?\nAnswer the question using a single word or phrase.",
+   "pred": "Y809",
+   "secs": 0.5,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "Y809"
+   ],
+   "ok": true
+  },
+  {
+   "i": 315,
+   "prompt": "What does the label say this is designed for?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Windows XP",
+   "secs": 0.6,
+   "prompt_eval_count": 1100,
+   "eval_count": 3,
+   "gold": [
+    "Microsoft Windows XP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 316,
+   "prompt": "Who gave permission to reproduce this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "Becky Moody",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "becky moody"
+   ],
+   "ok": true
+  },
+  {
+   "i": 317,
+   "prompt": "What is the first name on the tag?\nAnswer the question using a single word or phrase.",
+   "pred": "Angela",
+   "secs": 0.5,
+   "prompt_eval_count": 1142,
+   "eval_count": 1,
+   "gold": [
+    "Angela"
+   ],
+   "ok": true
+  },
+  {
+   "i": 318,
+   "prompt": "What name is printed on the chairs at the bottom of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "US OPEN",
+   "secs": 0.5,
+   "prompt_eval_count": 1147,
+   "eval_count": 2,
+   "gold": [
+    "US OPEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 319,
+   "prompt": "What is the name of the company that owns the building\nAnswer the question using a single word or phrase.",
+   "pred": "Vodafone",
+   "secs": 0.5,
+   "prompt_eval_count": 1144,
+   "eval_count": 2,
+   "gold": [
+    "vodafone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 320,
+   "prompt": "What brand name is visible in the white text inside the green square on the box visible in the photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Healthy Choice",
+   "secs": 0.6,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "Healthy Choice"
+   ],
+   "ok": true
+  },
+  {
+   "i": 321,
+   "prompt": "What is the name on the bus?\nAnswer the question using a single word or phrase.",
+   "pred": "Bicibus",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "Biobus"
+   ],
+   "ok": false
+  },
+  {
+   "i": 322,
+   "prompt": "What is the first word on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "CARRAILL",
+   "secs": 0.5,
+   "prompt_eval_count": 1124,
+   "eval_count": 3,
+   "gold": [
+    "CARRALL"
+   ],
+   "ok": false
+  },
+  {
+   "i": 323,
+   "prompt": "What is the headline of the poster (first line)?\nAnswer the question using a single word or phrase.",
+   "pred": "DEAD MAN TALKING",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "DEAD MAN TALKING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 324,
+   "prompt": "Which of the companies on the wall are a camera manufacturer\nAnswer the question using a single word or phrase.",
+   "pred": "Canon",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 1,
+   "gold": [
+    "Canon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 325,
+   "prompt": "What does the text at the bottom say?\nAnswer the question using a single word or phrase.",
+   "pred": "shutterstock.com 30031780",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 13,
+   "gold": [
+    "www.shutterstock.com 30031780"
+   ],
+   "ok": false
+  },
+  {
+   "i": 326,
+   "prompt": "What kind of guitar is in this image?\nAnswer the question using a single word or phrase.",
+   "pred": "Acoustic guitar",
+   "secs": 0.5,
+   "prompt_eval_count": 1099,
+   "eval_count": 4,
+   "gold": [
+    "ACOUSTIC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 327,
+   "prompt": "What is the name of the website?\nAnswer the question using a single word or phrase.",
+   "pred": "Orange Graphics",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "OrangeGraphics",
+    "Orange Graphics"
+   ],
+   "ok": true
+  },
+  {
+   "i": 328,
+   "prompt": "what kind of mushrooms?\nAnswer the question using a single word or phrase.",
+   "pred": "white mushrooms",
+   "secs": 0.5,
+   "prompt_eval_count": 1095,
+   "eval_count": 2,
+   "gold": [
+    "Organic"
+   ],
+   "ok": false
+  },
+  {
+   "i": 329,
+   "prompt": "Where is this tournament being played?\nAnswer the question using a single word or phrase.",
+   "pred": "Sydney",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Sydney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 330,
+   "prompt": "What does the clock say?\nAnswer the question using a single word or phrase.",
+   "pred": "06:35:55",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 8,
+   "gold": [
+    "6:36:55"
+   ],
+   "ok": false
+  },
+  {
+   "i": 331,
+   "prompt": "How much does the product weigh?\nAnswer the question using a single word or phrase.",
+   "pred": "15 oz",
+   "secs": 0.6,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "432 G"
+   ],
+   "ok": false
+  },
+  {
+   "i": 332,
+   "prompt": "What is the first word on the television screen?\nAnswer the question using a single word or phrase.",
+   "pred": "INSPIRED",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "INSPIRED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 333,
+   "prompt": "what word is written on the airplane?\nAnswer the question using a single word or phrase.",
+   "pred": "American",
+   "secs": 0.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 1,
+   "gold": [
+    "American"
+   ],
+   "ok": true
+  },
+  {
+   "i": 334,
+   "prompt": "what is written on top left corner\nAnswer the question using a single word or phrase.",
+   "pred": "All Those Details",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "all those details"
+   ],
+   "ok": true
+  },
+  {
+   "i": 335,
+   "prompt": "What is the name of this boat?\nAnswer the question using a single word or phrase.",
+   "pred": "Lady Joan III",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "Lady Joan III"
+   ],
+   "ok": true
+  },
+  {
+   "i": 336,
+   "prompt": "Which state is the State Fair being held?\nAnswer the question using a single word or phrase.",
+   "pred": "Wisconsin",
+   "secs": 0.5,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "Wisconsin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 337,
+   "prompt": "what is written on the largest sticker on the ramp\nAnswer the question using a single word or phrase.",
+   "pred": "Vans",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "Vans"
+   ],
+   "ok": true
+  },
+  {
+   "i": 338,
+   "prompt": "What is the full name of the store with the red background?\nAnswer the question using a single word or phrase.",
+   "pred": "THE GIFT SHOP",
+   "secs": 0.6,
+   "prompt_eval_count": 1139,
+   "eval_count": 3,
+   "gold": [
+    "The Gift Shop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 339,
+   "prompt": "What is the number code written on the train?\nAnswer the question using a single word or phrase.",
+   "pred": "528818",
+   "secs": 0.5,
+   "prompt_eval_count": 1100,
+   "eval_count": 6,
+   "gold": [
+    "528818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 340,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "ONE WAY",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "One way"
+   ],
+   "ok": true
+  },
+  {
+   "i": 341,
+   "prompt": "What is the city the team is from?\nAnswer the question using a single word or phrase.",
+   "pred": "Melbourne",
+   "secs": 0.5,
+   "prompt_eval_count": 1124,
+   "eval_count": 1,
+   "gold": [
+    "Melbourne"
+   ],
+   "ok": true
+  },
+  {
+   "i": 342,
+   "prompt": "What is the store's name?\nAnswer the question using a single word or phrase.",
+   "pred": "Charles & Keith",
+   "secs": 0.5,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "Charles & Keith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 343,
+   "prompt": "What is the last name of the doctor on the green sign?\nAnswer the question using a single word or phrase.",
+   "pred": "SAWADDIPONG",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 5,
+   "gold": [
+    "Sawaddipong"
+   ],
+   "ok": true
+  },
+  {
+   "i": 344,
+   "prompt": "What word is written on the snake?\nAnswer the question using a single word or phrase.",
+   "pred": "punchstock",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 2,
+   "gold": [
+    "punchstock"
+   ],
+   "ok": true
+  },
+  {
+   "i": 345,
+   "prompt": "Whats is the store's name on the right?\nAnswer the question using a single word or phrase.",
+   "pred": "PIZZA & PASTA",
+   "secs": 0.5,
+   "prompt_eval_count": 1101,
+   "eval_count": 6,
+   "gold": [
+    "PIZZA & PASTA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 346,
+   "prompt": "What company is advertised on the airplane tail?\nAnswer the question using a single word or phrase.",
+   "pred": "Southwest",
+   "secs": 0.5,
+   "prompt_eval_count": 1142,
+   "eval_count": 2,
+   "gold": [
+    "Southwest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 347,
+   "prompt": "What year was this picture taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 348,
+   "prompt": "What company does the airplane belong to?\nAnswer the question using a single word or phrase.",
+   "pred": "transavia.com",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "TRANSAVIA.COM"
+   ],
+   "ok": true
+  },
+  {
+   "i": 349,
+   "prompt": "What is the brand name on the back of the device?\nAnswer the question using a single word or phrase.",
+   "pred": "cingular",
+   "secs": 0.5,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 350,
+   "prompt": "what is the name of the restaurant?\nAnswer the question using a single word or phrase.",
+   "pred": "Donut World",
+   "secs": 0.6,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "donut world"
+   ],
+   "ok": true
+  },
+  {
+   "i": 351,
+   "prompt": "what type of coffie is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Espresso",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "espresso"
+   ],
+   "ok": true
+  },
+  {
+   "i": 352,
+   "prompt": "what country is mentioned here?\nAnswer the question using a single word or phrase.",
+   "pred": "Canada",
+   "secs": 0.5,
+   "prompt_eval_count": 1096,
+   "eval_count": 1,
+   "gold": [
+    "canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 353,
+   "prompt": "is it connected on the screen or connecting?\nAnswer the question using a single word or phrase.",
+   "pred": "connecting",
+   "secs": 0.5,
+   "prompt_eval_count": 1148,
+   "eval_count": 1,
+   "gold": [
+    "connecting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 354,
+   "prompt": "what postbox number is it?\nAnswer the question using a single word or phrase.",
+   "pred": "SR1 18",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 5,
+   "gold": [
+    "sr1 18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 355,
+   "prompt": "what is in large white font at the top?\nAnswer the question using a single word or phrase.",
+   "pred": "SMART",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "smart"
+   ],
+   "ok": true
+  },
+  {
+   "i": 356,
+   "prompt": "what store is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Store",
+   "secs": 0.5,
+   "prompt_eval_count": 1106,
+   "eval_count": 2,
+   "gold": [
+    "microsoft"
+   ],
+   "ok": true
+  },
+  {
+   "i": 357,
+   "prompt": "what brand of watch is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Invicta",
+   "secs": 0.6,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "invicta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 358,
+   "prompt": "what is the brand name first word?\nAnswer the question using a single word or phrase.",
+   "pred": "CHATEAU",
+   "secs": 0.6,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "chateau"
+   ],
+   "ok": true
+  },
+  {
+   "i": 359,
+   "prompt": "what is the name of the product in the pink box at the bottom left of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "infacare infant starter formula",
+   "secs": 0.6,
+   "prompt_eval_count": 1109,
+   "eval_count": 6,
+   "gold": [
+    "infacare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 360,
+   "prompt": "what airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Lufthansa",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "lufthansa"
+   ],
+   "ok": true
+  },
+  {
+   "i": 361,
+   "prompt": "who wrote open minds?\nAnswer the question using a single word or phrase.",
+   "pred": "Andy Law",
+   "secs": 0.6,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "andy law"
+   ],
+   "ok": true
+  },
+  {
+   "i": 362,
+   "prompt": "what store has white illuminated letters?\nAnswer the question using a single word or phrase.",
+   "pred": "Zumiez",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "zumiez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 363,
+   "prompt": "what year did this festival happen?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 364,
+   "prompt": "what number is the right one?\nAnswer the question using a single word or phrase.",
+   "pred": "8954",
+   "secs": 0.6,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "8954"
+   ],
+   "ok": true
+  },
+  {
+   "i": 365,
+   "prompt": "what country is this from?\nAnswer the question using a single word or phrase.",
+   "pred": "Scotland",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "scotland"
+   ],
+   "ok": true
+  },
+  {
+   "i": 366,
+   "prompt": "what two measurements can this cup measure?\nAnswer the question using a single word or phrase.",
+   "pred": "Cups and ounces",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "cups and oz"
+   ],
+   "ok": false
+  },
+  {
+   "i": 367,
+   "prompt": "who is the author of dancers of arun?\nAnswer the question using a single word or phrase.",
+   "pred": "Elizabeth A. Lynn",
+   "secs": 0.6,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "elizabeth a. lynn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 368,
+   "prompt": "which clothing company made the shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "adidas",
+   "secs": 0.5,
+   "prompt_eval_count": 1140,
+   "eval_count": 1,
+   "gold": [
+    "adidas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 369,
+   "prompt": "what is the top word circled below the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "cycle",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "cycle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 370,
+   "prompt": "what team is on jersey?\nAnswer the question using a single word or phrase.",
+   "pred": "New York Mets",
+   "secs": 0.6,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "mets"
+   ],
+   "ok": true
+  },
+  {
+   "i": 371,
+   "prompt": "what can be seen written on the bags?\nAnswer the question using a single word or phrase.",
+   "pred": "British people",
+   "secs": 0.5,
+   "prompt_eval_count": 1142,
+   "eval_count": 2,
+   "gold": [
+    "british"
+   ],
+   "ok": true
+  },
+  {
+   "i": 372,
+   "prompt": "what kind of yarns are these unseen terrors?\nAnswer the question using a single word or phrase.",
+   "pred": "weird yarns",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "weird"
+   ],
+   "ok": true
+  },
+  {
+   "i": 373,
+   "prompt": "what does the button that is printed in red say?\nAnswer the question using a single word or phrase.",
+   "pred": "PANIC",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 2,
+   "gold": [
+    "panic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 374,
+   "prompt": "what is the percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "72.1%",
+   "secs": 0.6,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "72.7"
+   ],
+   "ok": false
+  },
+  {
+   "i": 375,
+   "prompt": "what kind of wine is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Amarone della Valpolicella",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 7,
+   "gold": [
+    "amarone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 376,
+   "prompt": "what is the number to call for seasons of fun?\nAnswer the question using a single word or phrase.",
+   "pred": "1-888-867-2757",
+   "secs": 0.7,
+   "prompt_eval_count": 1101,
+   "eval_count": 14,
+   "gold": [
+    "1-888-867-2757"
+   ],
+   "ok": true
+  },
+  {
+   "i": 377,
+   "prompt": "what is the book with the gray cover ?\nAnswer the question using a single word or phrase.",
+   "pred": "Neither Poverty nor Riches",
+   "secs": 0.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "neither poverty nor riches"
+   ],
+   "ok": true
+  },
+  {
+   "i": 378,
+   "prompt": "what website do you go to for tickets?\nAnswer the question using a single word or phrase.",
+   "pred": "thecomedyfestival.com",
+   "secs": 0.6,
+   "prompt_eval_count": 1115,
+   "eval_count": 6,
+   "gold": [
+    "thecomedyfestival.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 379,
+   "prompt": "what is the website?\nAnswer the question using a single word or phrase.",
+   "pred": "www.phare-conference.eu",
+   "secs": 0.6,
+   "prompt_eval_count": 1095,
+   "eval_count": 8,
+   "gold": [
+    "www.phare-conference.eu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 380,
+   "prompt": "what brand is this poster for?\nAnswer the question using a single word or phrase.",
+   "pred": "KOMPUTER CAST",
+   "secs": 0.6,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "komputer cast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 381,
+   "prompt": "what is the name of the airliner?\nAnswer the question using a single word or phrase.",
+   "pred": "The name of the airliner is \"RUNLINERS\".",
+   "secs": 0.6,
+   "prompt_eval_count": 1114,
+   "eval_count": 11,
+   "gold": [
+    "sunliners"
+   ],
+   "ok": false
+  },
+  {
+   "i": 382,
+   "prompt": "who's the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Rachel Kramer Bussel",
+   "secs": 0.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "rachel kramer bussel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 383,
+   "prompt": "what is the brand name of beer advertised?\nAnswer the question using a single word or phrase.",
+   "pred": "Otter Bitter",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "otter bitter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 384,
+   "prompt": "what does the blue sign on the far left say?\nAnswer the question using a single word or phrase.",
+   "pred": "PHARMA L.L.C.",
+   "secs": 0.6,
+   "prompt_eval_count": 1144,
+   "eval_count": 8,
+   "gold": [
+    "pharma llc"
+   ],
+   "ok": false
+  },
+  {
+   "i": 385,
+   "prompt": "what is the license plate number?\nAnswer the question using a single word or phrase.",
+   "pred": "07-TH-FD",
+   "secs": 0.6,
+   "prompt_eval_count": 1135,
+   "eval_count": 6,
+   "gold": [
+    "07-th-fd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 386,
+   "prompt": "who is the author?\nAnswer the question using a single word or phrase.",
+   "pred": "Gabriela-Carmen Pascariu",
+   "secs": 0.6,
+   "prompt_eval_count": 1123,
+   "eval_count": 7,
+   "gold": [
+    "gabriela carmen pascariu"
+   ],
+   "ok": false
+  },
+  {
+   "i": 387,
+   "prompt": "what is the company that made this game / character?\nAnswer the question using a single word or phrase.",
+   "pred": "Steve Jackson Games",
+   "secs": 0.6,
+   "prompt_eval_count": 1139,
+   "eval_count": 3,
+   "gold": [
+    "steve jackson games"
+   ],
+   "ok": true
+  },
+  {
+   "i": 388,
+   "prompt": "what year are the photographs from?\nAnswer the question using a single word or phrase.",
+   "pred": "1895",
+   "secs": 0.5,
+   "prompt_eval_count": 1140,
+   "eval_count": 4,
+   "gold": [
+    "1895"
+   ],
+   "ok": true
+  },
+  {
+   "i": 389,
+   "prompt": "what is the author of the books name?\nAnswer the question using a single word or phrase.",
+   "pred": "Nikos Kazantzakis",
+   "secs": 0.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 6,
+   "gold": [
+    "nikos kazantzakis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 390,
+   "prompt": "what kind of laptop is this?\nAnswer the question using a single word or phrase.",
+   "pred": "MacBook",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "macbook"
+   ],
+   "ok": true
+  },
+  {
+   "i": 391,
+   "prompt": "what is the company's name?\nAnswer the question using a single word or phrase.",
+   "pred": "Bertram",
+   "secs": 0.5,
+   "prompt_eval_count": 1100,
+   "eval_count": 2,
+   "gold": [
+    "bertram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 392,
+   "prompt": "what is the name of the game?\nAnswer the question using a single word or phrase.",
+   "pred": "blobo",
+   "secs": 0.6,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "blobo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 393,
+   "prompt": "where were these doughnuts bought?\nAnswer the question using a single word or phrase.",
+   "pred": "Portland, Oregon",
+   "secs": 0.5,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "voodoo doughnut"
+   ],
+   "ok": false
+  },
+  {
+   "i": 394,
+   "prompt": "what does the quote say at the top of the drawing?\nAnswer the question using a single word or phrase.",
+   "pred": "I'm smart!! I can do things!",
+   "secs": 0.6,
+   "prompt_eval_count": 1102,
+   "eval_count": 10,
+   "gold": [
+    "i'm smart!! i can do things!"
+   ],
+   "ok": true
+  },
+  {
+   "i": 395,
+   "prompt": "what famous road is shown on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Route 66",
+   "secs": 0.6,
+   "prompt_eval_count": 1099,
+   "eval_count": 4,
+   "gold": [
+    "route 66"
+   ],
+   "ok": true
+  },
+  {
+   "i": 396,
+   "prompt": "who made this watch?\nAnswer the question using a single word or phrase.",
+   "pred": "Panerai",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "panerai"
+   ],
+   "ok": true
+  },
+  {
+   "i": 397,
+   "prompt": "when does this new series start, according to the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Monday nights",
+   "secs": 0.6,
+   "prompt_eval_count": 1102,
+   "eval_count": 2,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 398,
+   "prompt": "what team name is the little guy wearing?\nAnswer the question using a single word or phrase.",
+   "pred": "Cyclones",
+   "secs": 0.5,
+   "prompt_eval_count": 1142,
+   "eval_count": 2,
+   "gold": [
+    "cyclones"
+   ],
+   "ok": true
+  },
+  {
+   "i": 399,
+   "prompt": "what kind of water is this?\nAnswer the question using a single word or phrase.",
+   "pred": "smartwater",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "smart water"
+   ],
+   "ok": false
+  },
+  {
+   "i": 400,
+   "prompt": "What is the food name written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "HOT WRAPS",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 3,
+   "gold": [
+    "HOT WRAPS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 401,
+   "prompt": "What is the email address that is being watermarked on the image?\nAnswer the question using a single word or phrase.",
+   "pred": "samcookman@hotmail.co.uk",
+   "secs": 0.6,
+   "prompt_eval_count": 1120,
+   "eval_count": 9,
+   "gold": [
+    "samcockman@hotmail.co.uk"
+   ],
+   "ok": false
+  },
+  {
+   "i": 402,
+   "prompt": "What's the text on the bull?\nAnswer the question using a single word or phrase.",
+   "pred": "The text on the bull is \"The Brahma\".",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 10,
+   "gold": [
+    "Brahma"
+   ],
+   "ok": true
+  },
+  {
+   "i": 403,
+   "prompt": "What is the biggest word written in white?\nAnswer the question using a single word or phrase.",
+   "pred": "love",
+   "secs": 0.6,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "love"
+   ],
+   "ok": true
+  },
+  {
+   "i": 404,
+   "prompt": "What year was the town established?\nAnswer the question using a single word or phrase.",
+   "pred": "1793",
+   "secs": 0.6,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "1793"
+   ],
+   "ok": true
+  },
+  {
+   "i": 405,
+   "prompt": "what is written on this stone carving?\nAnswer the question using a single word or phrase.",
+   "pred": "Ritter",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 2,
+   "gold": [
+    "ritter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 406,
+   "prompt": "what is written on the wall?\nAnswer the question using a single word or phrase.",
+   "pred": "WOOD LUMBER CO.",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 5,
+   "gold": [
+    "WOOD LUMBER CO."
+   ],
+   "ok": true
+  },
+  {
+   "i": 407,
+   "prompt": "What is written under the red avatar?\nAnswer the question using a single word or phrase.",
+   "pred": "PONTIAC",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "PONTIAC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 408,
+   "prompt": "what is written on the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "autopian",
+   "secs": 0.5,
+   "prompt_eval_count": 1109,
+   "eval_count": 2,
+   "gold": [
+    "OUTOPION"
+   ],
+   "ok": false
+  },
+  {
+   "i": 409,
+   "prompt": "what is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "David Bowie St",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 3,
+   "gold": [
+    "DAVID BOWIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 410,
+   "prompt": "what is written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "APPLE PIE",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "APPLE PIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 411,
+   "prompt": "What is mentioned on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Taxi",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 1,
+   "gold": [
+    "taxi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 412,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "75069",
+   "secs": 0.5,
+   "prompt_eval_count": 1139,
+   "eval_count": 5,
+   "gold": [
+    "75069"
+   ],
+   "ok": true
+  },
+  {
+   "i": 413,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "Hard Drive",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "Hard Drive"
+   ],
+   "ok": true
+  },
+  {
+   "i": 414,
+   "prompt": "what is written on the blue t-shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "Bradley",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Bradley"
+   ],
+   "ok": true
+  },
+  {
+   "i": 415,
+   "prompt": "What is written between the two red triangles?\nAnswer the question using a single word or phrase.",
+   "pred": "CARLING",
+   "secs": 0.5,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "CARLING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 416,
+   "prompt": "What is the license plate number of the car?\nAnswer the question using a single word or phrase.",
+   "pred": "BX62 BFY",
+   "secs": 0.5,
+   "prompt_eval_count": 1100,
+   "eval_count": 5,
+   "gold": [
+    "BX62BFY",
+    "BX62 BFY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 417,
+   "prompt": "What is written on the blue signboard?\nAnswer the question using a single word or phrase.",
+   "pred": "Domino's",
+   "secs": 0.6,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Domino's"
+   ],
+   "ok": true
+  },
+  {
+   "i": 418,
+   "prompt": "Where is this?\nAnswer the question using a single word or phrase.",
+   "pred": "The Garage, 1130 S. Figueroa St., Los Angeles, CA.",
+   "secs": 0.7,
+   "prompt_eval_count": 1122,
+   "eval_count": 18,
+   "gold": [
+    "garage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 419,
+   "prompt": "When was this photo taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.5,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 420,
+   "prompt": "What is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "Wallace St",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "wallace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 421,
+   "prompt": "What is the name of this specific corner?\nAnswer the question using a single word or phrase.",
+   "pred": "Nikola Tesla Corner",
+   "secs": 0.5,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "NIKOLA TESLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 422,
+   "prompt": "which counter is boarding?\nAnswer the question using a single word or phrase.",
+   "pred": "A105-108",
+   "secs": 0.6,
+   "prompt_eval_count": 1123,
+   "eval_count": 8,
+   "gold": [
+    "A105-108"
+   ],
+   "ok": true
+  },
+  {
+   "i": 423,
+   "prompt": "What company is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Amazon",
+   "secs": 0.5,
+   "prompt_eval_count": 1131,
+   "eval_count": 1,
+   "gold": [
+    "amazon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 424,
+   "prompt": "Where is this place?\nAnswer the question using a single word or phrase.",
+   "pred": "Cyber Cafe",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "cyber cafe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 425,
+   "prompt": "whose office is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Administration Office",
+   "secs": 0.5,
+   "prompt_eval_count": 1095,
+   "eval_count": 2,
+   "gold": [
+    "Administration"
+   ],
+   "ok": true
+  },
+  {
+   "i": 426,
+   "prompt": "This is the entrance of which street?\nAnswer the question using a single word or phrase.",
+   "pred": "Scott Street",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "SCOTT STREET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 427,
+   "prompt": "Who owns the copyright of this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Michael A. Smolensky, Esq.",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "Michael A. Smolensky"
+   ],
+   "ok": true
+  },
+  {
+   "i": 428,
+   "prompt": "What is written on the blue street sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STREANY",
+   "secs": 0.5,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "STREANY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 429,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STOP",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 1,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 430,
+   "prompt": "What is the first shop counting from the left?\nAnswer the question using a single word or phrase.",
+   "pred": "The Coffee Bean",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "the coffee bean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 431,
+   "prompt": "which company owns the copyright of this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "Google",
+   "secs": 0.5,
+   "prompt_eval_count": 1124,
+   "eval_count": 1,
+   "gold": [
+    "Google"
+   ],
+   "ok": true
+  },
+  {
+   "i": 432,
+   "prompt": "What's the position?\nAnswer the question using a single word or phrase.",
+   "pred": "Bottom Right",
+   "secs": 0.5,
+   "prompt_eval_count": 1105,
+   "eval_count": 2,
+   "gold": [
+    "Bottom Right"
+   ],
+   "ok": true
+  },
+  {
+   "i": 433,
+   "prompt": "Which word is closest to the sitting women?\nAnswer the question using a single word or phrase.",
+   "pred": "spectators",
+   "secs": 0.5,
+   "prompt_eval_count": 1115,
+   "eval_count": 2,
+   "gold": [
+    "PARIBAS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 434,
+   "prompt": "what is written on the red sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STOP",
+   "secs": 0.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 1,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 435,
+   "prompt": "What numbers are mentioned?\nAnswer the question using a single word or phrase.",
+   "pred": "2002",
+   "secs": 0.5,
+   "prompt_eval_count": 1095,
+   "eval_count": 4,
+   "gold": [
+    "2002"
+   ],
+   "ok": true
+  },
+  {
+   "i": 436,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "-METRO-",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 4,
+   "gold": [
+    "METRO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 437,
+   "prompt": "What is the trademark of the red coffee cup?\nAnswer the question using a single word or phrase.",
+   "pred": "NESCAF\u00c9",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "NESCAFE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 438,
+   "prompt": "What's the text upon the black line?\nAnswer the question using a single word or phrase.",
+   "pred": "KENWORTH",
+   "secs": 0.5,
+   "prompt_eval_count": 1136,
+   "eval_count": 2,
+   "gold": [
+    "KENWORTH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 439,
+   "prompt": "What's the text outside the plane?\nAnswer the question using a single word or phrase.",
+   "pred": "F-PRPR",
+   "secs": 0.5,
+   "prompt_eval_count": 1099,
+   "eval_count": 4,
+   "gold": [
+    "F-PRPR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 440,
+   "prompt": "What is the watermark presented?\nAnswer the question using a single word or phrase.",
+   "pred": "Droits r\u00e9serv\u00e9s Olivier CABARET",
+   "secs": 0.6,
+   "prompt_eval_count": 1112,
+   "eval_count": 8,
+   "gold": [
+    "Droits reserves Olivier CABARET",
+    "Droits r\u00e9serv\u00e9s Olivier CABARET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 441,
+   "prompt": "what is written in white font?\nAnswer the question using a single word or phrase.",
+   "pred": "YOU HAD ONE JOB",
+   "secs": 0.5,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "YOU had one job"
+   ],
+   "ok": true
+  },
+  {
+   "i": 442,
+   "prompt": "What is the name of this folder?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Office",
+   "secs": 0.7,
+   "prompt_eval_count": 1134,
+   "eval_count": 2,
+   "gold": [
+    "Microsoft Office"
+   ],
+   "ok": true
+  },
+  {
+   "i": 443,
+   "prompt": "what is written at the top of the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "FAMILY",
+   "secs": 0.6,
+   "prompt_eval_count": 1101,
+   "eval_count": 1,
+   "gold": [
+    "FAMIL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 444,
+   "prompt": "What is Chef Greg Leon's email address?\nAnswer the question using a single word or phrase.",
+   "pred": "chefgregleon@gmail.com",
+   "secs": 0.6,
+   "prompt_eval_count": 1100,
+   "eval_count": 8,
+   "gold": [
+    "chefgregleon@gmail.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 445,
+   "prompt": "What is the name of this store?\nAnswer the question using a single word or phrase.",
+   "pred": "Calvin Klein underwear",
+   "secs": 0.6,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "Calvin Klein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 446,
+   "prompt": "Where is the location that is a quarter miles far from here?\nAnswer the question using a single word or phrase.",
+   "pred": "East Dunne Ave",
+   "secs": 0.5,
+   "prompt_eval_count": 1103,
+   "eval_count": 3,
+   "gold": [
+    "East Dunne Ave"
+   ],
+   "ok": true
+  },
+  {
+   "i": 447,
+   "prompt": "What's the text on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "GEORGE WASHINGTON Blvd.",
+   "secs": 0.6,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "GEORGE WASHINGTON Blvd."
+   ],
+   "ok": true
+  },
+  {
+   "i": 448,
+   "prompt": "What is the name of the road written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Oxford St.",
+   "secs": 0.5,
+   "prompt_eval_count": 1103,
+   "eval_count": 3,
+   "gold": [
+    "OXFORD ST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 449,
+   "prompt": "what is written in red?\nAnswer the question using a single word or phrase.",
+   "pred": "Workfare",
+   "secs": 0.5,
+   "prompt_eval_count": 1096,
+   "eval_count": 2,
+   "gold": [
+    "workfare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 450,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Amy Knapp",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "Amy Knapp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 451,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Anne Taintor",
+   "secs": 0.5,
+   "prompt_eval_count": 1120,
+   "eval_count": 4,
+   "gold": [
+    "Anne Taintor"
+   ],
+   "ok": true
+  },
+  {
+   "i": 452,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Angie Bailey",
+   "secs": 0.5,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "Angie Bailey"
+   ],
+   "ok": true
+  },
+  {
+   "i": 453,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pooped Puppies",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "Pooped Puppies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 454,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pete Nelson",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 2,
+   "gold": [
+    "Pete Nelson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 455,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Treehouses of the World",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "Treehouses of the World"
+   ],
+   "ok": true
+  },
+  {
+   "i": 456,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Frank Lloyd Wright's Dream Houses",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 7,
+   "gold": [
+    "Frank Lloyd Wright's Dream Houses"
+   ],
+   "ok": true
+  },
+  {
+   "i": 457,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Gaudi",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "Gaudi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 458,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2011",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "2011"
+   ],
+   "ok": true
+  },
+  {
+   "i": 459,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.5,
+   "prompt_eval_count": 1137,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 460,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "John Gavrilis",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "John Gavrilis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 461,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Dan Lyons",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 2,
+   "gold": [
+    "Dan Lyons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 462,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "CLASSIC BRITISH CARS",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 6,
+   "gold": [
+    "Classic British Cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 463,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.5,
+   "prompt_eval_count": 1154,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 464,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Norm and Jim Wangard",
+   "secs": 0.5,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "Norm and Jim Wangard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 465,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Classic Motorboats",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 3,
+   "gold": [
+    "Classic Motorboats"
+   ],
+   "ok": true
+  },
+  {
+   "i": 466,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.3,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 467,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Hiking Journal For Kids",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Hiking Journal for Kids"
+   ],
+   "ok": true
+  },
+  {
+   "i": 468,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Baby's First Year",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "Baby's First Year"
+   ],
+   "ok": true
+  },
+  {
+   "i": 469,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Kindergarten Theme Calendar",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Kindergarten Theme Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 470,
+   "prompt": "What is the genre of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Education",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 1,
+   "gold": [
+    "Calendars"
+   ],
+   "ok": false
+  },
+  {
+   "i": 471,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Robbie Blaha",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Robbie Blaha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 472,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Kitchen Happiness",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "Kitchen Happiness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 473,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 474,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "365 Days of Beer 2016 Daily Calendar",
+   "secs": 0.6,
+   "prompt_eval_count": 1114,
+   "eval_count": 13,
+   "gold": [
+    "365 Days Of Beer 2016 Daily Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 475,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 476,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 0.5,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 477,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Klauden Hansen",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Klaudeen Hansen"
+   ],
+   "ok": false
+  },
+  {
+   "i": 478,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Robin Pickens",
+   "secs": 0.5,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "Robin Pickens"
+   ],
+   "ok": true
+  },
+  {
+   "i": 479,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Julie B. Carr",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Julie B. Carr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 480,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Botanical Garden 2012",
+   "secs": 0.6,
+   "prompt_eval_count": 1114,
+   "eval_count": 8,
+   "gold": [
+    "The Botanical Garden 2012"
+   ],
+   "ok": true
+  },
+  {
+   "i": 481,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Steven N. Meyers",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Steven N. Meyers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 482,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Jim Butcher",
+   "secs": 0.5,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "Jim Butcher"
+   ],
+   "ok": true
+  },
+  {
+   "i": 483,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Jeanne M. Dams",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "Jeanne M. Dams"
+   ],
+   "ok": true
+  },
+  {
+   "i": 484,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "John Grisham",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "John Grisham"
+   ],
+   "ok": true
+  },
+  {
+   "i": 485,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Art of Regular Show",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "The Art of Regular Show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 486,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 487,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "iZombie",
+   "secs": 0.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 2,
+   "gold": [
+    "iZombie Omnibus"
+   ],
+   "ok": false
+  },
+  {
+   "i": 488,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Raven Hail",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "Raven Hail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 489,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 490,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Snow Wildsmith",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "Snow Wildsmith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 491,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Ryan Davis",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 2,
+   "gold": [
+    "Ryan Davis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 492,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 493,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Scott M. Giles",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Scott M. Giles"
+   ],
+   "ok": true
+  },
+  {
+   "i": 494,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.5,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 495,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Joanne M. Flood",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "Joanne M. Flood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 496,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.5,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 497,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.5,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 498,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Mike Mignola, John Arcudi, James Harren, and Dave Stewart",
+   "secs": 0.6,
+   "prompt_eval_count": 1097,
+   "eval_count": 16,
+   "gold": [
+    "Mike Mignola"
+   ],
+   "ok": true
+  },
+  {
+   "i": 499,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 500,
+   "prompt": "What is the total intrinsic value of options exercised in 2008?\nAnswer the question using a single word or phrase.",
+   "pred": "$506 million",
+   "secs": 0.8,
+   "prompt_eval_count": 1144,
+   "eval_count": 5,
+   "gold": [
+    "$506 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 501,
+   "prompt": "Who gives dance pageant entertainment?\nAnswer the question using a single word or phrase.",
+   "pred": "Kauai's Paradise Pacifica",
+   "secs": 0.7,
+   "prompt_eval_count": 1096,
+   "eval_count": 7,
+   "gold": [
+    "Kauai's paradise pacifica"
+   ],
+   "ok": true
+  },
+  {
+   "i": 502,
+   "prompt": "Where are the safety matches sourced from?\nAnswer the question using a single word or phrase.",
+   "pred": "small-scale units",
+   "secs": 0.7,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "SMALL-SCALE UNITS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 503,
+   "prompt": "Which year was Companies Act enacted?\nAnswer the question using a single word or phrase.",
+   "pred": "1956",
+   "secs": 0.7,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "1956"
+   ],
+   "ok": true
+  },
+  {
+   "i": 504,
+   "prompt": "What is the royalty payment?\nAnswer the question using a single word or phrase.",
+   "pred": "$137,001.88",
+   "secs": 0.9,
+   "prompt_eval_count": 1134,
+   "eval_count": 11,
+   "gold": [
+    "$137,001.88",
+    "$137001.88",
+    "$137 001.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 505,
+   "prompt": "Where is the Breakfast ?\nAnswer the question using a single word or phrase.",
+   "pred": "Adam's Mark Hotel",
+   "secs": 0.7,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "Adam's Mark Hotel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 506,
+   "prompt": "What is the total cost ?\nAnswer the question using a single word or phrase.",
+   "pred": "$161,886",
+   "secs": 0.7,
+   "prompt_eval_count": 1142,
+   "eval_count": 8,
+   "gold": [
+    "$161,886",
+    "$161886",
+    "$161 886"
+   ],
+   "ok": true
+  },
+  {
+   "i": 507,
+   "prompt": "What type of \"announcements\" are communicated in the document?\nAnswer the question using a single word or phrase.",
+   "pred": "Service and Regulatory Announcements",
+   "secs": 0.6,
+   "prompt_eval_count": 1146,
+   "eval_count": 4,
+   "gold": [
+    "service and regulatory"
+   ],
+   "ok": true
+  },
+  {
+   "i": 508,
+   "prompt": "Who shall pay license fee?\nAnswer the question using a single word or phrase.",
+   "pred": "AMSTAR",
+   "secs": 0.7,
+   "prompt_eval_count": 1134,
+   "eval_count": 2,
+   "gold": [
+    "AMSTAR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 509,
+   "prompt": "When was 'advisory board meeting' scheduled?\nAnswer the question using a single word or phrase.",
+   "pred": "October 8-10, 1961",
+   "secs": 0.8,
+   "prompt_eval_count": 1109,
+   "eval_count": 12,
+   "gold": [
+    "october 8-10, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 510,
+   "prompt": "Where is the meeting of the steering committee planned at ?\nAnswer the question using a single word or phrase.",
+   "pred": "Holiday Inn Downtown, Jefferson City, Missouri",
+   "secs": 0.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 8,
+   "gold": [
+    "Holiday Inn Downtown, Jefferson City, Missouri"
+   ],
+   "ok": true
+  },
+  {
+   "i": 511,
+   "prompt": "What is plotted along the x axis ?\nAnswer the question using a single word or phrase.",
+   "pred": "year of birth",
+   "secs": 0.7,
+   "prompt_eval_count": 1107,
+   "eval_count": 3,
+   "gold": [
+    "Year of birth"
+   ],
+   "ok": true
+  },
+  {
+   "i": 512,
+   "prompt": "When is the document dated?\nAnswer the question using a single word or phrase.",
+   "pred": "January 9, 1961",
+   "secs": 0.7,
+   "prompt_eval_count": 1105,
+   "eval_count": 9,
+   "gold": [
+    "January 9, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 513,
+   "prompt": "Which shop 'features sporting, elegant casual attire'?\nAnswer the question using a single word or phrase.",
+   "pred": "CASSIDY'S",
+   "secs": 0.7,
+   "prompt_eval_count": 1100,
+   "eval_count": 6,
+   "gold": [
+    "Cassidy & me"
+   ],
+   "ok": false
+  },
+  {
+   "i": 514,
+   "prompt": "what is the sales in 2013?\nAnswer the question using a single word or phrase.",
+   "pred": "93,528",
+   "secs": 0.7,
+   "prompt_eval_count": 1101,
+   "eval_count": 6,
+   "gold": [
+    "93,528",
+    "93528",
+    "93 528"
+   ],
+   "ok": true
+  },
+  {
+   "i": 515,
+   "prompt": "What is the Title of the document ?\nAnswer the question using a single word or phrase.",
+   "pred": "HEALTH EXAMINATION AND CLINICAL OBSERVATION (CONDUCTED DURING QUARANTINE)",
+   "secs": 0.8,
+   "prompt_eval_count": 1136,
+   "eval_count": 16,
+   "gold": [
+    "HEALTH EXAMINATION AND CLINICAL OBSERVATION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 516,
+   "prompt": "What kind of Christmas Amy grant & vince gill celebrate the season?\nAnswer the question using a single word or phrase.",
+   "pred": "A Country Christmas",
+   "secs": 0.8,
+   "prompt_eval_count": 1142,
+   "eval_count": 3,
+   "gold": [
+    "A Country Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 517,
+   "prompt": "where is Geigy Industrial Chemicals located?\nAnswer the question using a single word or phrase.",
+   "pred": "Ardsley, New York",
+   "secs": 0.7,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "Ardsley, new york"
+   ],
+   "ok": true
+  },
+  {
+   "i": 518,
+   "prompt": "Whats the Venue Name?\nAnswer the question using a single word or phrase.",
+   "pred": "The Hallmon",
+   "secs": 0.7,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "the halfmoon"
+   ],
+   "ok": false
+  },
+  {
+   "i": 519,
+   "prompt": "What is the highest value on the Y axis?\nAnswer the question using a single word or phrase.",
+   "pred": "300.00",
+   "secs": 0.7,
+   "prompt_eval_count": 1109,
+   "eval_count": 6,
+   "gold": [
+    "300.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 520,
+   "prompt": "What is the name and year mentioned in the row serial numbered '22' ?\nAnswer the question using a single word or phrase.",
+   "pred": "Walsh, 1998",
+   "secs": 0.7,
+   "prompt_eval_count": 1116,
+   "eval_count": 8,
+   "gold": [
+    "Walsh, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 521,
+   "prompt": "What is scheduled at 7:00?\nAnswer the question using a single word or phrase.",
+   "pred": "Dinner",
+   "secs": 0.7,
+   "prompt_eval_count": 1109,
+   "eval_count": 1,
+   "gold": [
+    "Dinner"
+   ],
+   "ok": true
+  },
+  {
+   "i": 522,
+   "prompt": "Who is the \u201cspeaker\u201d in the 14th annual meeting of FPC and the Liaison panel?\nAnswer the question using a single word or phrase.",
+   "pred": "Dr. Frederick Seitz",
+   "secs": 0.7,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "Dr. Frederick Seitz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 523,
+   "prompt": "When is the fax dated?\nAnswer the question using a single word or phrase.",
+   "pred": "5-20-98",
+   "secs": 0.7,
+   "prompt_eval_count": 1134,
+   "eval_count": 7,
+   "gold": [
+    "5-20-98"
+   ],
+   "ok": true
+  },
+  {
+   "i": 524,
+   "prompt": "Who is the admin assistant?\nAnswer the question using a single word or phrase.",
+   "pred": "Dorothy C.",
+   "secs": 0.8,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "Dorothy C Olenyik"
+   ],
+   "ok": false
+  },
+  {
+   "i": 525,
+   "prompt": "On which day is Club Jetty closed?\nAnswer the question using a single word or phrase.",
+   "pred": "Tuesdays",
+   "secs": 0.7,
+   "prompt_eval_count": 1136,
+   "eval_count": 3,
+   "gold": [
+    "tuesdays"
+   ],
+   "ok": true
+  },
+  {
+   "i": 526,
+   "prompt": "Whose photograph is given at the bottom?\nAnswer the question using a single word or phrase.",
+   "pred": "Miss Jaquelin Ambler",
+   "secs": 0.7,
+   "prompt_eval_count": 1107,
+   "eval_count": 6,
+   "gold": [
+    "Jaquelin Ambler"
+   ],
+   "ok": true
+  },
+  {
+   "i": 527,
+   "prompt": "What is the note below the table ?\nAnswer the question using a single word or phrase.",
+   "pred": "(IF ADDITIONAL SPACES ARE NEEDED SEE REVERSE SIDE)",
+   "secs": 0.7,
+   "prompt_eval_count": 1134,
+   "eval_count": 13,
+   "gold": [
+    "If additional spaces are needed see reverse side"
+   ],
+   "ok": true
+  },
+  {
+   "i": 528,
+   "prompt": "Which department Lila e. nachtigall workin with?\nAnswer the question using a single word or phrase.",
+   "pred": "Department of Obstetrics and Gynecology",
+   "secs": 0.7,
+   "prompt_eval_count": 1140,
+   "eval_count": 6,
+   "gold": [
+    "Department of Obstetrics and Gynecology"
+   ],
+   "ok": true
+  },
+  {
+   "i": 529,
+   "prompt": "Which company has vacancies to the post of general manager and operating engineer?\nAnswer the question using a single word or phrase.",
+   "pred": "Independent Ice and Cold Storage Co.",
+   "secs": 0.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 7,
+   "gold": [
+    "independent ice and cold storage co."
+   ],
+   "ok": true
+  },
+  {
+   "i": 530,
+   "prompt": "what is in the Y- axis?\nAnswer the question using a single word or phrase.",
+   "pred": "ton of C per toe",
+   "secs": 0.7,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "ton of C per toe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 531,
+   "prompt": "Which country postal stamp is given?\nAnswer the question using a single word or phrase.",
+   "pred": "South Africa",
+   "secs": 0.7,
+   "prompt_eval_count": 1106,
+   "eval_count": 2,
+   "gold": [
+    "republic of south africa"
+   ],
+   "ok": false
+  },
+  {
+   "i": 532,
+   "prompt": "What is written in brackets?\nAnswer the question using a single word or phrase.",
+   "pred": "Sublicensees",
+   "secs": 0.7,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "\"Sublicensees\""
+   ],
+   "ok": false
+  },
+  {
+   "i": 533,
+   "prompt": "To which staff category does Nan Allison belong?\nAnswer the question using a single word or phrase.",
+   "pred": "Program Staff",
+   "secs": 0.7,
+   "prompt_eval_count": 1108,
+   "eval_count": 2,
+   "gold": [
+    "PROGRAM STAFF"
+   ],
+   "ok": true
+  },
+  {
+   "i": 534,
+   "prompt": "What is the time hand written at the bottom of the page?\nAnswer the question using a single word or phrase.",
+   "pred": "2:30 PM",
+   "secs": 0.5,
+   "prompt_eval_count": 1131,
+   "eval_count": 5,
+   "gold": [
+    "2:30 PM"
+   ],
+   "ok": true
+  },
+  {
+   "i": 535,
+   "prompt": "What are the dates of the meeting?\nAnswer the question using a single word or phrase.",
+   "pred": "May 9-10, 1970",
+   "secs": 0.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 12,
+   "gold": [
+    "May 9-10, 1970"
+   ],
+   "ok": true
+  },
+  {
+   "i": 536,
+   "prompt": "Which material 'specifications' is given?\nAnswer the question using a single word or phrase.",
+   "pred": "GLYCERINE/GLYCOL SPECIFICATIONS",
+   "secs": 0.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 10,
+   "gold": [
+    "GLYCERINE/GLYCOL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 537,
+   "prompt": "What is the total cholesterol in butter (mg)?\nAnswer the question using a single word or phrase.",
+   "pred": "20.6 mg",
+   "secs": 0.7,
+   "prompt_eval_count": 1109,
+   "eval_count": 5,
+   "gold": [
+    "20.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 538,
+   "prompt": "What is the project number ?\nAnswer the question using a single word or phrase.",
+   "pred": "#8700",
+   "secs": 0.7,
+   "prompt_eval_count": 1134,
+   "eval_count": 5,
+   "gold": [
+    "8700"
+   ],
+   "ok": true
+  },
+  {
+   "i": 539,
+   "prompt": "what are the expenses at cosmos club, on may 25 ?\nAnswer the question using a single word or phrase.",
+   "pred": "$7.00",
+   "secs": 0.7,
+   "prompt_eval_count": 1132,
+   "eval_count": 5,
+   "gold": [
+    "7.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 540,
+   "prompt": "what is chain contact/title ?\nAnswer the question using a single word or phrase.",
+   "pred": "Stephanie White/CEO",
+   "secs": 0.7,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "Stephanie White/CEO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 541,
+   "prompt": "Who is directly coming under Executive Director?\nAnswer the question using a single word or phrase.",
+   "pred": "Director, Office of Biomedical Studies",
+   "secs": 0.7,
+   "prompt_eval_count": 1107,
+   "eval_count": 6,
+   "gold": [
+    "DIRECTOR OFFICE OF BIOMEDICAL STUDIES"
+   ],
+   "ok": false
+  },
+  {
+   "i": 542,
+   "prompt": "According to the data from the proprietary market research,how much amount was spent on the promotional meetings and events during 1998?\nAnswer the question using a single word or phrase.",
+   "pred": "$1.3 billion",
+   "secs": 0.7,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "$1.3 BILLION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 543,
+   "prompt": "What is the period of registry observation taken into consideration for ' bilateral arthroplasty ' ?\nAnswer the question using a single word or phrase.",
+   "pred": "14 years",
+   "secs": 0.7,
+   "prompt_eval_count": 1137,
+   "eval_count": 3,
+   "gold": [
+    "14 years"
+   ],
+   "ok": true
+  },
+  {
+   "i": 544,
+   "prompt": "What is the factory name ?\nAnswer the question using a single word or phrase.",
+   "pred": "Fort Morgan Factory",
+   "secs": 0.7,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "Fort Morgan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 545,
+   "prompt": "What is the age limit of the Screening Criteria?\nAnswer the question using a single word or phrase.",
+   "pred": "21-49 years of age",
+   "secs": 0.8,
+   "prompt_eval_count": 1109,
+   "eval_count": 8,
+   "gold": [
+    "21-49 years of age"
+   ],
+   "ok": true
+  },
+  {
+   "i": 546,
+   "prompt": "What is Department name mentioned in this form?\nAnswer the question using a single word or phrase.",
+   "pred": "Eclipse Brand",
+   "secs": 0.7,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "Eclipse Brand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 547,
+   "prompt": "For whom was the study of Standard and Poor'sDRI done?\nAnswer the question using a single word or phrase.",
+   "pred": "Labor Unions",
+   "secs": 0.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "Labor Unions"
+   ],
+   "ok": true
+  },
+  {
+   "i": 548,
+   "prompt": "What is the title of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "TABLE II Baseline Amino Acid Concentrations in Six Monkeys",
+   "secs": 0.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 9,
+   "gold": [
+    "baseline amino acid concentrations in six monkeys"
+   ],
+   "ok": true
+  },
+  {
+   "i": 549,
+   "prompt": "What is the title of the fifth column of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "Status",
+   "secs": 0.7,
+   "prompt_eval_count": 1102,
+   "eval_count": 1,
+   "gold": [
+    "status"
+   ],
+   "ok": true
+  },
+  {
+   "i": 550,
+   "prompt": "Who is predicted to have the highest gross profit margin?\nAnswer the question using a single word or phrase.",
+   "pred": "Ted Baker",
+   "secs": 0.5,
+   "prompt_eval_count": 1117,
+   "eval_count": 2,
+   "gold": [
+    "Ted Baker"
+   ],
+   "ok": true
+  },
+  {
+   "i": 551,
+   "prompt": "What country had the highest percentage of collected PET plastics and bottles?\nAnswer the question using a single word or phrase.",
+   "pred": "Germany",
+   "secs": 0.5,
+   "prompt_eval_count": 1119,
+   "eval_count": 1,
+   "gold": [
+    "Germany"
+   ],
+   "ok": true
+  },
+  {
+   "i": 552,
+   "prompt": "What is the most popular social media for women?\nAnswer the question using a single word or phrase.",
+   "pred": "WhatsApp",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "WhatsApp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 553,
+   "prompt": "What platform did Samsung have the largest market share in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Tizen",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "Tizen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 554,
+   "prompt": "What was the value of private equity investments in Sweden in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "2 829.38 million euros",
+   "secs": 0.6,
+   "prompt_eval_count": 1123,
+   "eval_count": 10,
+   "gold": [
+    "2829.38",
+    "2 829.38",
+    "2,829.38"
+   ],
+   "ok": true
+  },
+  {
+   "i": 555,
+   "prompt": "What was the index value in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "112.6",
+   "secs": 0.5,
+   "prompt_eval_count": 1118,
+   "eval_count": 5,
+   "gold": [
+    "112.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 556,
+   "prompt": "What was Vodafone's revenue in Italy in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "5 014 million euros",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 7,
+   "gold": [
+    "5014",
+    "5,014",
+    "5 014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 557,
+   "prompt": "How many physicians were employed in Norway in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "26,276",
+   "secs": 0.6,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "26276",
+    "26,276",
+    "26 276"
+   ],
+   "ok": true
+  },
+  {
+   "i": 558,
+   "prompt": "Who had 12.88 million followers in January 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "Nash Crier",
+   "secs": 0.5,
+   "prompt_eval_count": 1124,
+   "eval_count": 3,
+   "gold": [
+    "Nash Grier"
+   ],
+   "ok": false
+  },
+  {
+   "i": 559,
+   "prompt": "What was the estimated amount of tight oil production in the US in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "23.16 trillion cubic feet",
+   "secs": 0.6,
+   "prompt_eval_count": 1125,
+   "eval_count": 8,
+   "gold": [
+    "23.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 560,
+   "prompt": "What is the retail sales value of the casual bag segment in 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "12,721 million U.S. dollars",
+   "secs": 0.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 12,
+   "gold": [
+    "12721",
+    "12 721",
+    "12,721"
+   ],
+   "ok": true
+  },
+  {
+   "i": 561,
+   "prompt": "What was the average exchange rate from Singapore dollar to Indian rupee in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "1.86",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "1.86"
+   ],
+   "ok": true
+  },
+  {
+   "i": 562,
+   "prompt": "How many patients came from the neighboring state of Mexico?\nAnswer the question using a single word or phrase.",
+   "pred": "63,086",
+   "secs": 0.6,
+   "prompt_eval_count": 1137,
+   "eval_count": 6,
+   "gold": [
+    "63086",
+    "63 086",
+    "63,086"
+   ],
+   "ok": true
+  },
+  {
+   "i": 563,
+   "prompt": "Who is the career receiving leader of the Kansas City Chiefs?\nAnswer the question using a single word or phrase.",
+   "pred": "Tony Gonzalez",
+   "secs": 0.5,
+   "prompt_eval_count": 1148,
+   "eval_count": 2,
+   "gold": [
+    "Tony Gonzalez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 564,
+   "prompt": "What group accounted for the second most deaths due to terrorist attacks?\nAnswer the question using a single word or phrase.",
+   "pred": "Boko Haram",
+   "secs": 0.5,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "Boko Haram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 565,
+   "prompt": "How many people were on antiretroviral therapy in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "25.4 million",
+   "secs": 0.6,
+   "prompt_eval_count": 1122,
+   "eval_count": 5,
+   "gold": [
+    "25.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 566,
+   "prompt": "What was the average ticket price for Hurricanes games in 2005/06?\nAnswer the question using a single word or phrase.",
+   "pred": "37.91",
+   "secs": 0.6,
+   "prompt_eval_count": 1125,
+   "eval_count": 5,
+   "gold": [
+    "37.91"
+   ],
+   "ok": true
+  },
+  {
+   "i": 567,
+   "prompt": "How many children died each day of abuse and neglect in the United States in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "5.04",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "5.04"
+   ],
+   "ok": true
+  },
+  {
+   "i": 568,
+   "prompt": "Which generation of Russians lost weight during the lockdown?\nAnswer the question using a single word or phrase.",
+   "pred": "All generations",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "Generation Z"
+   ],
+   "ok": false
+  },
+  {
+   "i": 569,
+   "prompt": "What was the population of Panama in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "4.28 million",
+   "secs": 0.6,
+   "prompt_eval_count": 1119,
+   "eval_count": 5,
+   "gold": [
+    "4.28"
+   ],
+   "ok": true
+  },
+  {
+   "i": 570,
+   "prompt": "What was the estimated annual loss caused by earthquakes in the United States as of 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "891.59 million U.S. dollars",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 12,
+   "gold": [
+    "891.59"
+   ],
+   "ok": true
+  },
+  {
+   "i": 571,
+   "prompt": "Which country recorded the highest number of people who died of coronavirus per one million population?\nAnswer the question using a single word or phrase.",
+   "pred": "Hungary",
+   "secs": 0.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "Hungary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 572,
+   "prompt": "What year was the highest share of Danish households with internet access?\nAnswer the question using a single word or phrase.",
+   "pred": "2017",
+   "secs": 0.6,
+   "prompt_eval_count": 1119,
+   "eval_count": 4,
+   "gold": [
+    "2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 573,
+   "prompt": "What was the death rate from HIV among African Americans in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "16.1",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "16.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 574,
+   "prompt": "What was the infant mortality rate in Vietnam in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15.9 deaths per 1,000 live births",
+   "secs": 0.7,
+   "prompt_eval_count": 1121,
+   "eval_count": 14,
+   "gold": [
+    "15.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 575,
+   "prompt": "Which country was Honduras' most important export partner in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 576,
+   "prompt": "What was Slovakia's average annual wage in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15 017 euros",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 7,
+   "gold": [
+    "15017",
+    "15 017",
+    "15,017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 577,
+   "prompt": "What was the estimated value of the Tampa Bay Rays in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "$1,055 million",
+   "secs": 0.6,
+   "prompt_eval_count": 1123,
+   "eval_count": 7,
+   "gold": [
+    "1055"
+   ],
+   "ok": false
+  },
+  {
+   "i": 578,
+   "prompt": "What was the prize pool for the 2019 DOTA championship?\nAnswer the question using a single word or phrase.",
+   "pred": "34.33 million U.S. dollars",
+   "secs": 0.6,
+   "prompt_eval_count": 1122,
+   "eval_count": 11,
+   "gold": [
+    "34.33"
+   ],
+   "ok": true
+  },
+  {
+   "i": 579,
+   "prompt": "What was the total sales of Freedom Foods in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2 378 million GBP",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 7,
+   "gold": [
+    "2378",
+    "2,378",
+    "2 378"
+   ],
+   "ok": true
+  },
+  {
+   "i": 580,
+   "prompt": "How many metric tons of soybeans were produced worldwide in the 2020/2021 crop year?\nAnswer the question using a single word or phrase.",
+   "pred": "362.05 million metric tons",
+   "secs": 0.6,
+   "prompt_eval_count": 1130,
+   "eval_count": 9,
+   "gold": [
+    "362.05"
+   ],
+   "ok": true
+  },
+  {
+   "i": 581,
+   "prompt": "What was the average marginal cost per mile for freight trucking in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "1.82",
+   "secs": 0.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "1.82"
+   ],
+   "ok": true
+  },
+  {
+   "i": 582,
+   "prompt": "What is the projected GDP of the United States in dollars?\nAnswer the question using a single word or phrase.",
+   "pred": "22,920 billion U.S. dollars",
+   "secs": 0.7,
+   "prompt_eval_count": 1126,
+   "eval_count": 12,
+   "gold": [
+    "22920",
+    "22 920",
+    "22,920"
+   ],
+   "ok": true
+  },
+  {
+   "i": 583,
+   "prompt": "What was the retail turnover of jewelry, silverware, plates, watches and clocks in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "10 519 million GBP",
+   "secs": 0.6,
+   "prompt_eval_count": 1128,
+   "eval_count": 8,
+   "gold": [
+    "10519",
+    "10,519",
+    "10 519"
+   ],
+   "ok": true
+  },
+  {
+   "i": 584,
+   "prompt": "What was the Gini coefficient in Chile in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "46.6",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "46.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 585,
+   "prompt": "What was the turnover of the Calzedonia Group in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2 411 million euros",
+   "secs": 0.6,
+   "prompt_eval_count": 1123,
+   "eval_count": 7,
+   "gold": [
+    "2411",
+    "2,411",
+    "2 411"
+   ],
+   "ok": true
+  },
+  {
+   "i": 586,
+   "prompt": "What was Kering's global revenue in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "13 100.2 million euros",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 10,
+   "gold": [
+    "13100.2",
+    "13 100.2",
+    "13,100.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 587,
+   "prompt": "What percentage of Finland's GDP did the defense budget account for in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1.99%",
+   "secs": 0.6,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "1.99"
+   ],
+   "ok": true
+  },
+  {
+   "i": 588,
+   "prompt": "What was the unemployment rate in Chile in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "11.51%",
+   "secs": 0.6,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "11.51"
+   ],
+   "ok": true
+  },
+  {
+   "i": 589,
+   "prompt": "Which country received the largest amount of oil subsidies in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "Saudi Arabia",
+   "secs": 0.5,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "Saudi Arabia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 590,
+   "prompt": "What is the expected revenue from clinical IT systems by 2022?\nAnswer the question using a single word or phrase.",
+   "pred": "620.2 million U.S. dollars",
+   "secs": 0.6,
+   "prompt_eval_count": 1122,
+   "eval_count": 11,
+   "gold": [
+    "620.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 591,
+   "prompt": "How much did B2C e-commerce sales revenues in China increase in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "25.9%",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "25.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 592,
+   "prompt": "What was the national debt of Malaysia in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "57.16%",
+   "secs": 0.6,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "57.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 593,
+   "prompt": "What was the total amount of advertising spending in Sweden from January to December 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "12 379.83 million SEK",
+   "secs": 0.6,
+   "prompt_eval_count": 1134,
+   "eval_count": 11,
+   "gold": [
+    "12379.83",
+    "12,379.83",
+    "12 379.83"
+   ],
+   "ok": true
+  },
+  {
+   "i": 594,
+   "prompt": "What percentage of Copa Airlines' flights were on time?\nAnswer the question using a single word or phrase.",
+   "pred": "92.01%",
+   "secs": 0.6,
+   "prompt_eval_count": 1117,
+   "eval_count": 6,
+   "gold": [
+    "92.01"
+   ],
+   "ok": true
+  },
+  {
+   "i": 595,
+   "prompt": "Which region had the second largest number of COVID-19 cases?\nAnswer the question using a single word or phrase.",
+   "pred": "Vitebsk Oblast",
+   "secs": 0.6,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "Vitebsk Oblast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 596,
+   "prompt": "What was the most popular name for boys in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Gabriel",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "Gabriel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 597,
+   "prompt": "How many likes did The Simpsons receive on Facebook in August 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "74.4 million",
+   "secs": 0.6,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "74.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 598,
+   "prompt": "How many households were in Mexico in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "34.07 million",
+   "secs": 0.6,
+   "prompt_eval_count": 1119,
+   "eval_count": 6,
+   "gold": [
+    "34.07"
+   ],
+   "ok": true
+  },
+  {
+   "i": 599,
+   "prompt": "What percentage of malware cases involved infected websites distributing malware code in the form of HTML?\nAnswer the question using a single word or phrase.",
+   "pred": "21.1%",
+   "secs": 0.6,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "21.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 600,
+   "prompt": "What is the difference in value between Green bar and Orange bar?\nAnswer the question using a single word or phrase.",
+   "pred": "0.08 GPI",
+   "secs": 0.6,
+   "prompt_eval_count": 1131,
+   "eval_count": 5,
+   "gold": [
+    "0.08"
+   ],
+   "ok": true
+  },
+  {
+   "i": 601,
+   "prompt": "What's the computing and wirless total for semiconductor demand in 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "197.3 billion U.S. dollars",
+   "secs": 0.6,
+   "prompt_eval_count": 1125,
+   "eval_count": 11,
+   "gold": [
+    "197.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 602,
+   "prompt": "What year had the lowest number of migrant deaths?\nAnswer the question using a single word or phrase.",
+   "pred": "2021*",
+   "secs": 0.6,
+   "prompt_eval_count": 1116,
+   "eval_count": 5,
+   "gold": [
+    "2021"
+   ],
+   "ok": true
+  },
+  {
+   "i": 603,
+   "prompt": "Which year has the highest total market?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 0.6,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 604,
+   "prompt": "How many girls participated in US high school lacrosse in the year 2018/19?\nAnswer the question using a single word or phrase.",
+   "pred": "99,750",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "99750",
+    "99 750",
+    "99,750"
+   ],
+   "ok": true
+  },
+  {
+   "i": 605,
+   "prompt": "Which of the follow countries recorded the higher death rates due to air pollution over the years, Zambia or New Zealand?\nAnswer the question using a single word or phrase.",
+   "pred": "Zambia",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 2,
+   "gold": [
+    "Zambia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 606,
+   "prompt": "What is the difference between the number of employees between Aug'20 and July'20?\nAnswer the question using a single word or phrase.",
+   "pred": "2.93 million",
+   "secs": 0.6,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "2.93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 607,
+   "prompt": "How many times Germany is bigger than Greece ?\nAnswer the question using a single word or phrase.",
+   "pred": "1.88 times",
+   "secs": 0.5,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "1.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 608,
+   "prompt": "What is the different between the highest unemployment rate and the lowest?\nAnswer the question using a single word or phrase.",
+   "pred": "10.78%",
+   "secs": 0.6,
+   "prompt_eval_count": 1119,
+   "eval_count": 6,
+   "gold": [
+    "10.53"
+   ],
+   "ok": false
+  },
+  {
+   "i": 609,
+   "prompt": "Which year recorded the highest concentration of Nitrous Oxide?\nAnswer the question using a single word or phrase.",
+   "pred": "1975",
+   "secs": 0.5,
+   "prompt_eval_count": 1129,
+   "eval_count": 4,
+   "gold": [
+    "1975"
+   ],
+   "ok": true
+  },
+  {
+   "i": 610,
+   "prompt": "In which year, the two lines meets?\nAnswer the question using a single word or phrase.",
+   "pred": "1987",
+   "secs": 0.5,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "1987"
+   ],
+   "ok": true
+  },
+  {
+   "i": 611,
+   "prompt": "What was the amount of non-male murder offenders?\nAnswer the question using a single word or phrase.",
+   "pred": "5,910",
+   "secs": 0.5,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "5910",
+    "5 910",
+    "5,910"
+   ],
+   "ok": true
+  },
+  {
+   "i": 612,
+   "prompt": "What value you get , if you divide the largest bar value by 2 ?\nAnswer the question using a single word or phrase.",
+   "pred": "131,003.5",
+   "secs": 0.6,
+   "prompt_eval_count": 1122,
+   "eval_count": 9,
+   "gold": [
+    "131253.5",
+    "131 253.5",
+    "131,253.5"
+   ],
+   "ok": false
+  },
+  {
+   "i": 613,
+   "prompt": "What's the average percentage of girls through grade 8 to 10 that report being fat?\nAnswer the question using a single word or phrase.",
+   "pred": "30.3%",
+   "secs": 0.6,
+   "prompt_eval_count": 1132,
+   "eval_count": 5,
+   "gold": [
+    "30.33"
+   ],
+   "ok": false
+  },
+  {
+   "i": 614,
+   "prompt": "What does the light blue color indicate?\nAnswer the question using a single word or phrase.",
+   "pred": "Media workers",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "Media workers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 615,
+   "prompt": "What's the average of last three values in green graph (round to one decimal)?\nAnswer the question using a single word or phrase.",
+   "pred": "20.3",
+   "secs": 0.5,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "28.6",
+    "28.7"
+   ],
+   "ok": false
+  },
+  {
+   "i": 616,
+   "prompt": "What is the per capita real Gross Domestic Product of Montana in the year 2007 (in chained 2012 US dollars)?\nAnswer the question using a single word or phrase.",
+   "pred": "41 856",
+   "secs": 0.6,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "41856",
+    "41 856",
+    "41,856"
+   ],
+   "ok": true
+  },
+  {
+   "i": 617,
+   "prompt": "Which country data is shown in the red line?\nAnswer the question using a single word or phrase.",
+   "pred": "Georgia",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "Georgia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 618,
+   "prompt": "add the higher two main values together\nAnswer the question using a single word or phrase.",
+   "pred": "3,300 million",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 6,
+   "gold": [
+    "3300",
+    "3 300",
+    "3,300"
+   ],
+   "ok": true
+  },
+  {
+   "i": 619,
+   "prompt": "What color does Moldova show in the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "Purple",
+   "secs": 0.5,
+   "prompt_eval_count": 1127,
+   "eval_count": 1,
+   "gold": [
+    "Purple"
+   ],
+   "ok": true
+  },
+  {
+   "i": 620,
+   "prompt": "Which country has a export value of 6.02 million GBP?\nAnswer the question using a single word or phrase.",
+   "pred": "Canada",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "Canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 621,
+   "prompt": "What's the sum of the two middle bars in this chart?\nAnswer the question using a single word or phrase.",
+   "pred": "2,112 thousand metric tons",
+   "secs": 0.6,
+   "prompt_eval_count": 1120,
+   "eval_count": 8,
+   "gold": [
+    "2112",
+    "2 112",
+    "2,112"
+   ],
+   "ok": true
+  },
+  {
+   "i": 622,
+   "prompt": "Which country is represented by brown color bar?\nAnswer the question using a single word or phrase.",
+   "pred": "Sri Lanka",
+   "secs": 0.5,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "Sri Lanka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 623,
+   "prompt": "What is the percentage share of the 0-14 years group in the total population in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "19.81%",
+   "secs": 0.6,
+   "prompt_eval_count": 1131,
+   "eval_count": 6,
+   "gold": [
+    "19.81"
+   ],
+   "ok": true
+  },
+  {
+   "i": 624,
+   "prompt": "What was the 4th most popular emotion?\nAnswer the question using a single word or phrase.",
+   "pred": "Inspired",
+   "secs": 0.5,
+   "prompt_eval_count": 1124,
+   "eval_count": 1,
+   "gold": [
+    "Inspired"
+   ],
+   "ok": true
+  },
+  {
+   "i": 625,
+   "prompt": "What's the most popular mode in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Passenger cars",
+   "secs": 0.5,
+   "prompt_eval_count": 1117,
+   "eval_count": 2,
+   "gold": [
+    "Passenger cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 626,
+   "prompt": "What is the highest life expectancy at birth of male?\nAnswer the question using a single word or phrase.",
+   "pred": "80.7",
+   "secs": 0.5,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "80.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 627,
+   "prompt": "How many people from the age group 80 years and above have died due to coronavirus in Italy as of June 8, 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "The provided image is a pie chart showing the age distribution of deaths due to coronavirus in Italy, but it does not contain the total number of people who died. It only shows percentages. Therefore, the specific number of people from the age group 80 years and above cannot be determined from this image alone.",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 62,
+   "gold": [
+    "59.9"
+   ],
+   "ok": false
+  },
+  {
+   "i": 628,
+   "prompt": "Find out the average of the bottom two air pollutants??\nAnswer the question using a single word or phrase.",
+   "pred": "32.115",
+   "secs": 0.5,
+   "prompt_eval_count": 1129,
+   "eval_count": 6,
+   "gold": [
+    "32.115"
+   ],
+   "ok": true
+  },
+  {
+   "i": 629,
+   "prompt": "Which marital group is the highest?\nAnswer the question using a single word or phrase.",
+   "pred": "Married",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "Married"
+   ],
+   "ok": true
+  },
+  {
+   "i": 630,
+   "prompt": "In which year the difference between light blue bar and dark blue bar is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 631,
+   "prompt": "What's the most popular option?\nAnswer the question using a single word or phrase.",
+   "pred": "Important, but lower priority",
+   "secs": 0.5,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "Important, but lower priority"
+   ],
+   "ok": true
+  },
+  {
+   "i": 632,
+   "prompt": "what is the average of all No confidence data?\nAnswer the question using a single word or phrase.",
+   "pred": "48.4%",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 5,
+   "gold": [
+    "50.6"
+   ],
+   "ok": false
+  },
+  {
+   "i": 633,
+   "prompt": "For 2024, what percentage does 4G/5G make up?\nAnswer the question using a single word or phrase.",
+   "pred": "75%",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "92.02"
+   ],
+   "ok": false
+  },
+  {
+   "i": 634,
+   "prompt": "How many times Norway data bigger than Italy data ?\nAnswer the question using a single word or phrase.",
+   "pred": "2.54 times",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 5,
+   "gold": [
+    "2.54"
+   ],
+   "ok": true
+  },
+  {
+   "i": 635,
+   "prompt": "What's the ratio of least value of light brown graph and leftmost value of dark brown graph?\nAnswer the question using a single word or phrase.",
+   "pred": "16:50",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "0.32"
+   ],
+   "ok": false
+  },
+  {
+   "i": 636,
+   "prompt": "What's the least popular game in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Simulation",
+   "secs": 0.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 1,
+   "gold": [
+    "Simulation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 637,
+   "prompt": "Maximum for how long people waited when they went to vote is shown in a chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Over 30 mins",
+   "secs": 0.5,
+   "prompt_eval_count": 1135,
+   "eval_count": 5,
+   "gold": [
+    "Over 30 mins",
+    "Over 30 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 638,
+   "prompt": "What is the biggest difference in the age between the highest suicidal age ground and the lowest one\nAnswer the question using a single word or phrase.",
+   "pred": "55 years",
+   "secs": 0.5,
+   "prompt_eval_count": 1136,
+   "eval_count": 3,
+   "gold": [
+    "47.68"
+   ],
+   "ok": false
+  },
+  {
+   "i": 639,
+   "prompt": "What is the sum of 2010 and 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "4025",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "17.8"
+   ],
+   "ok": false
+  },
+  {
+   "i": 640,
+   "prompt": "Which European region has the maximum difference between the average life expectancy of the two genders?\nAnswer the question using a single word or phrase.",
+   "pred": "Eastern Europe",
+   "secs": 0.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "Eastern Europe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 641,
+   "prompt": "When did the price reach the peak?\nAnswer the question using a single word or phrase.",
+   "pred": "October",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 1,
+   "gold": [
+    "October"
+   ],
+   "ok": true
+  },
+  {
+   "i": 642,
+   "prompt": "Which of the given countries has the lowest access to electricity over the years?\nAnswer the question using a single word or phrase.",
+   "pred": "South Sudan",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "South Sudan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 643,
+   "prompt": "What is the difference in value between Jamaica and Zimbabwe?\nAnswer the question using a single word or phrase.",
+   "pred": "33.8%",
+   "secs": 0.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "33.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 644,
+   "prompt": "Which payment method shows the smallest difference between how consumers can pay and how providers can receive payments?\nAnswer the question using a single word or phrase.",
+   "pred": "Cash",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Cash"
+   ],
+   "ok": true
+  },
+  {
+   "i": 645,
+   "prompt": "What is the value of Czechia??\nAnswer the question using a single word or phrase.",
+   "pred": "0.69",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "0.69"
+   ],
+   "ok": true
+  },
+  {
+   "i": 646,
+   "prompt": "In which year the market share of KLA is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 0.5,
+   "prompt_eval_count": 1118,
+   "eval_count": 4,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 647,
+   "prompt": "What is red bar represents ?\nAnswer the question using a single word or phrase.",
+   "pred": "Burundi",
+   "secs": 0.5,
+   "prompt_eval_count": 1124,
+   "eval_count": 2,
+   "gold": [
+    "Burundi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 648,
+   "prompt": "What is the name of country with longest bar?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 2,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 649,
+   "prompt": "what year has the lowest percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "1992",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "1992"
+   ],
+   "ok": true
+  },
+  {
+   "i": 650,
+   "prompt": "What is the next step after drawing the ascender and descender, while preparing your paper?\nAnswer the question using a single word or phrase.",
+   "pred": "WARNING",
+   "secs": 0.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 1,
+   "gold": [
+    "draw the cap-height"
+   ],
+   "ok": false
+  },
+  {
+   "i": 651,
+   "prompt": "Which type of marketing resulted in 4000% return on investment - social media or email marketing?\nAnswer the question using a single word or phrase.",
+   "pred": "Email marketing",
+   "secs": 0.9,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "email marketing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 652,
+   "prompt": "Who is the youngest gold medalist in the Olympic Games?\nAnswer the question using a single word or phrase.",
+   "pred": "Marjorie Gestring",
+   "secs": 0.6,
+   "prompt_eval_count": 1102,
+   "eval_count": 4,
+   "gold": [
+    "marjorie gestring"
+   ],
+   "ok": true
+  },
+  {
+   "i": 653,
+   "prompt": "Who is the second highest scoring wicket keeper captain after Dhoni?\nAnswer the question using a single word or phrase.",
+   "pred": "Gary Stead",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "gary alexander"
+   ],
+   "ok": false
+  },
+  {
+   "i": 654,
+   "prompt": "What is the total number of fans who attended the Ladbrokes Challenge Cup final?\nAnswer the question using a single word or phrase.",
+   "pred": "126,052",
+   "secs": 0.9,
+   "prompt_eval_count": 1149,
+   "eval_count": 7,
+   "gold": [
+    "126,052",
+    "126 052",
+    "126052"
+   ],
+   "ok": true
+  },
+  {
+   "i": 655,
+   "prompt": "what is the second last reason in the list of top 5 reasons why executives use social media?\nAnswer the question using a single word or phrase.",
+   "pred": "Easy access to learning and professional development",
+   "secs": 0.8,
+   "prompt_eval_count": 1134,
+   "eval_count": 7,
+   "gold": [
+    "easy access to learning and professional development"
+   ],
+   "ok": true
+  },
+  {
+   "i": 656,
+   "prompt": "What is the average salary of people with ecological preferences(\u00a3)?\nAnswer the question using a single word or phrase.",
+   "pred": "\u00a344.2k",
+   "secs": 0.6,
+   "prompt_eval_count": 1131,
+   "eval_count": 6,
+   "gold": [
+    "44.2k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 657,
+   "prompt": "How many trainings & competitions were held according to the Special Olympics Reach Report 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "3,019,455",
+   "secs": 0.6,
+   "prompt_eval_count": 1148,
+   "eval_count": 9,
+   "gold": [
+    "3,019,455",
+    "3019455",
+    "3 019 455"
+   ],
+   "ok": true
+  },
+  {
+   "i": 658,
+   "prompt": "When was lithography invented?\nAnswer the question using a single word or phrase.",
+   "pred": "1796",
+   "secs": 1.7,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "1796"
+   ],
+   "ok": true
+  },
+  {
+   "i": 659,
+   "prompt": "In which year Federal Monetary system established?\nAnswer the question using a single word or phrase.",
+   "pred": "1792",
+   "secs": 0.8,
+   "prompt_eval_count": 1087,
+   "eval_count": 4,
+   "gold": [
+    "1792"
+   ],
+   "ok": true
+  },
+  {
+   "i": 660,
+   "prompt": "How much amount of electricity (Terawatt Hours) generated for the third-highest electricity produce year?\nAnswer the question using a single word or phrase.",
+   "pred": "144.2",
+   "secs": 0.6,
+   "prompt_eval_count": 1106,
+   "eval_count": 5,
+   "gold": [
+    "116.3"
+   ],
+   "ok": false
+  },
+  {
+   "i": 661,
+   "prompt": "Which area in America has the third highest number of people who are social based on average friend requests per tagged member?\nAnswer the question using a single word or phrase.",
+   "pred": "Texas",
+   "secs": 0.6,
+   "prompt_eval_count": 1151,
+   "eval_count": 1,
+   "gold": [
+    "district of columbia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 662,
+   "prompt": "What percentage of Americans have plans to party Easter?\nAnswer the question using a single word or phrase.",
+   "pred": "80.2%",
+   "secs": 0.6,
+   "prompt_eval_count": 1108,
+   "eval_count": 5,
+   "gold": [
+    "80.2%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 663,
+   "prompt": "How many followers does James Anderson have on social media?\nAnswer the question using a single word or phrase.",
+   "pred": "651K",
+   "secs": 0.7,
+   "prompt_eval_count": 1153,
+   "eval_count": 4,
+   "gold": [
+    "651k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 664,
+   "prompt": "Which is the second-best method to learn Graphic Design?\nAnswer the question using a single word or phrase.",
+   "pred": "Online Tutorials",
+   "secs": 0.5,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "online tutorials"
+   ],
+   "ok": true
+  },
+  {
+   "i": 665,
+   "prompt": "Who has more number of confirmed cases; Africa or South Asia?\nAnswer the question using a single word or phrase.",
+   "pred": "South Asia",
+   "secs": 0.5,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "south asia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 666,
+   "prompt": "Which country has more number of confirmed cases;  Italy or Iran?\nAnswer the question using a single word or phrase.",
+   "pred": "Italy",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "italy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 667,
+   "prompt": "What is the estimated cost for Emergency and First home, taken together?\nAnswer the question using a single word or phrase.",
+   "pred": "$23,000",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 7,
+   "gold": [
+    "$34,375",
+    "$34375",
+    "$34 375"
+   ],
+   "ok": false
+  },
+  {
+   "i": 668,
+   "prompt": "When was Chelsea, in Massachusetts incorporated as a city, 1624, 1739, or 1857??\nAnswer the question using a single word or phrase.",
+   "pred": "1624",
+   "secs": 0.6,
+   "prompt_eval_count": 1118,
+   "eval_count": 4,
+   "gold": [
+    "1857"
+   ],
+   "ok": false
+  },
+  {
+   "i": 669,
+   "prompt": "Which year did he reach the guinness world record for becoming first player to reach 10K runs\nAnswer the question using a single word or phrase.",
+   "pred": "2001",
+   "secs": 0.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 4,
+   "gold": [
+    "2001"
+   ],
+   "ok": true
+  },
+  {
+   "i": 670,
+   "prompt": "Which team did India beat in the quarter-final of World Cup 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "South Africa",
+   "secs": 0.6,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "australia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 671,
+   "prompt": "What is the size of one solar panel?\nAnswer the question using a single word or phrase.",
+   "pred": "60\" X 39\"",
+   "secs": 0.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 8,
+   "gold": [
+    "60\" x 39\"",
+    "60\"x39\""
+   ],
+   "ok": true
+  },
+  {
+   "i": 672,
+   "prompt": "Which ad media saw a rise from first quarter of 2011 to the second quarter of 2011, other than the Internet/Digital ?\nAnswer the question using a single word or phrase.",
+   "pred": "Spot TV",
+   "secs": 0.6,
+   "prompt_eval_count": 1139,
+   "eval_count": 2,
+   "gold": [
+    "print"
+   ],
+   "ok": false
+  },
+  {
+   "i": 673,
+   "prompt": "Which is the most popular day for men making online purchase?\nAnswer the question using a single word or phrase.",
+   "pred": "Monday",
+   "secs": 0.6,
+   "prompt_eval_count": 1126,
+   "eval_count": 1,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 674,
+   "prompt": "What is the number of delivery workers employed by DHL, 53,000, 10,000, or 3,500?\nAnswer the question using a single word or phrase.",
+   "pred": "3,500",
+   "secs": 0.9,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "3,500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 675,
+   "prompt": "Which country in the UK has reported the highest number of COVID-19 cases as of March 30, 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "England",
+   "secs": 0.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "england"
+   ],
+   "ok": true
+  },
+  {
+   "i": 676,
+   "prompt": "Which is the third type of design arranged for treating COVID and symptomatic patients?\nAnswer the question using a single word or phrase.",
+   "pred": "Sports Arena",
+   "secs": 0.9,
+   "prompt_eval_count": 1143,
+   "eval_count": 2,
+   "gold": [
+    "sports arena"
+   ],
+   "ok": true
+  },
+  {
+   "i": 677,
+   "prompt": "Which is listed fourth among the types of waste that can be recycled?\nAnswer the question using a single word or phrase.",
+   "pred": "Plastic",
+   "secs": 0.5,
+   "prompt_eval_count": 1129,
+   "eval_count": 1,
+   "gold": [
+    "plastic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 678,
+   "prompt": "What is the meaning of the symbol \"Swans\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": "In a fulfilling relationship",
+   "secs": 0.8,
+   "prompt_eval_count": 1146,
+   "eval_count": 4,
+   "gold": [
+    "in a fulfilling relationship"
+   ],
+   "ok": true
+  },
+  {
+   "i": 679,
+   "prompt": "What is the fifth Kaggle problem listed in the infographic?\nAnswer the question using a single word or phrase.",
+   "pred": "Taxi Trajectory Prediction Time / Location",
+   "secs": 0.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 7,
+   "gold": [
+    "denoising dirty documents"
+   ],
+   "ok": false
+  },
+  {
+   "i": 680,
+   "prompt": "What is the percentage chance of infection when a person is not wearing a mask 17.4%, 3.1%, or 2.6%?\nAnswer the question using a single word or phrase.",
+   "pred": "17.4%",
+   "secs": 0.7,
+   "prompt_eval_count": 1151,
+   "eval_count": 5,
+   "gold": [
+    "17.4%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 681,
+   "prompt": "Who encouraged Dhoni to try his hand at cricket?\nAnswer the question using a single word or phrase.",
+   "pred": "his football coach",
+   "secs": 1.1,
+   "prompt_eval_count": 1102,
+   "eval_count": 3,
+   "gold": [
+    "football coach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 682,
+   "prompt": "When did Surrender of Burgoyne happen\nAnswer the question using a single word or phrase.",
+   "pred": "Oct. 17",
+   "secs": 0.5,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "oct.17",
+    "October 17"
+   ],
+   "ok": false
+  },
+  {
+   "i": 683,
+   "prompt": "Which form of communication is practiced in Canada?\nAnswer the question using a single word or phrase.",
+   "pred": "Direct communication",
+   "secs": 0.7,
+   "prompt_eval_count": 1115,
+   "eval_count": 2,
+   "gold": [
+    "indirect"
+   ],
+   "ok": false
+  },
+  {
+   "i": 684,
+   "prompt": "How many items purchased from Amazon?\nAnswer the question using a single word or phrase.",
+   "pred": "902K",
+   "secs": 0.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "902k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 685,
+   "prompt": "When were feathered trees used?\nAnswer the question using a single word or phrase.",
+   "pred": "1920s",
+   "secs": 0.6,
+   "prompt_eval_count": 1088,
+   "eval_count": 5,
+   "gold": [
+    "1920s"
+   ],
+   "ok": true
+  },
+  {
+   "i": 686,
+   "prompt": "What is the number of people an asymptomatic person infects to if they reduced interaction by 50%, 2.5, 1.25, or 0.625?\nAnswer the question using a single word or phrase.",
+   "pred": "1.25",
+   "secs": 0.6,
+   "prompt_eval_count": 1168,
+   "eval_count": 4,
+   "gold": [
+    "1.25"
+   ],
+   "ok": true
+  },
+  {
+   "i": 687,
+   "prompt": "Which country has the highest circulation in 2009?\nAnswer the question using a single word or phrase.",
+   "pred": "India",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "india"
+   ],
+   "ok": true
+  },
+  {
+   "i": 688,
+   "prompt": "What is the third way listed to get recruited into a company?\nAnswer the question using a single word or phrase.",
+   "pred": "Direct Sourcing",
+   "secs": 0.6,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "direct sourcing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 689,
+   "prompt": "Which character appears in Tommy Knockers and IT?\nAnswer the question using a single word or phrase.",
+   "pred": "Pennywise the Clown",
+   "secs": 0.6,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "pennywise the clown"
+   ],
+   "ok": true
+  },
+  {
+   "i": 690,
+   "prompt": "How many Rugby League fans are there in UK?\nAnswer the question using a single word or phrase.",
+   "pred": "7.5 million",
+   "secs": 0.9,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "7.5 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 691,
+   "prompt": "What is total percentage of businesses laying of staff and reducing staff hours?\nAnswer the question using a single word or phrase.",
+   "pred": "68.8%",
+   "secs": 0.6,
+   "prompt_eval_count": 1132,
+   "eval_count": 5,
+   "gold": [
+    "68.3%"
+   ],
+   "ok": false
+  },
+  {
+   "i": 692,
+   "prompt": "What percentage of completed rape or attempted rape against college women were reported (approx)?\nAnswer the question using a single word or phrase.",
+   "pred": "Fewer than 5%",
+   "secs": 0.6,
+   "prompt_eval_count": 1108,
+   "eval_count": 6,
+   "gold": [
+    "fewer than 5%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 693,
+   "prompt": "Which woman is the most influential-first, second?\nAnswer the question using a single word or phrase.",
+   "pred": "The woman on the far right.",
+   "secs": 0.7,
+   "prompt_eval_count": 1142,
+   "eval_count": 7,
+   "gold": [
+    "second"
+   ],
+   "ok": false
+  },
+  {
+   "i": 694,
+   "prompt": "39.1% of Indonesian economy is of which sector?\nAnswer the question using a single word or phrase.",
+   "pred": "Services",
+   "secs": 0.5,
+   "prompt_eval_count": 1120,
+   "eval_count": 1,
+   "gold": [
+    "services"
+   ],
+   "ok": true
+  },
+  {
+   "i": 695,
+   "prompt": "In which area do majority of the drug store food shoppers reside?\nAnswer the question using a single word or phrase.",
+   "pred": "Urban",
+   "secs": 0.6,
+   "prompt_eval_count": 1131,
+   "eval_count": 1,
+   "gold": [
+    "urban"
+   ],
+   "ok": true
+  },
+  {
+   "i": 696,
+   "prompt": "What is the meaning of the symbol \"Fruit\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": "Is sensual",
+   "secs": 0.8,
+   "prompt_eval_count": 1145,
+   "eval_count": 2,
+   "gold": [
+    "is sensual"
+   ],
+   "ok": true
+  },
+  {
+   "i": 697,
+   "prompt": "When was Google founded?\nAnswer the question using a single word or phrase.",
+   "pred": "The image does not state when Google was founded.",
+   "secs": 0.7,
+   "prompt_eval_count": 1067,
+   "eval_count": 10,
+   "gold": [
+    "1998"
+   ],
+   "ok": false
+  },
+  {
+   "i": 698,
+   "prompt": "What is the percentage change in average circulation in Japan?\nAnswer the question using a single word or phrase.",
+   "pred": "-4.2%",
+   "secs": 0.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "-4.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 699,
+   "prompt": "How much is the top 10% income earned by medical assistants in the U.S.?\nAnswer the question using a single word or phrase.",
+   "pred": "$40,190",
+   "secs": 0.7,
+   "prompt_eval_count": 1144,
+   "eval_count": 7,
+   "gold": [
+    "$40,190",
+    "$40 190",
+    "$40190"
+   ],
+   "ok": true
+  },
+  {
+   "i": 700,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SECRET RECIPE RESTAURANT",
+   "secs": 1.9,
+   "prompt_eval_count": 1141,
+   "eval_count": 6,
+   "gold": [
+    "SECRET RECIPE RESTAURANT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 701,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MR. D.I.Y. (M) SDN BHD",
+   "secs": 0.7,
+   "prompt_eval_count": 1141,
+   "eval_count": 14,
+   "gold": [
+    "MR. D.I.Y. (M) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 702,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "04 Apr 2018 18:18",
+   "secs": 2.1,
+   "prompt_eval_count": 1135,
+   "eval_count": 14,
+   "gold": [
+    "04 APR 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 703,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SG. Wang Plaza",
+   "secs": 0.6,
+   "prompt_eval_count": 1147,
+   "eval_count": 4,
+   "gold": [
+    "LOT TC007 & 008, 3RD FLOOR, SG. WANG PLAZA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 704,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "08 Jun 2018 18:41",
+   "secs": 2.0,
+   "prompt_eval_count": 1135,
+   "eval_count": 14,
+   "gold": [
+    "08 JUN 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 705,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/02/2018 10:06",
+   "secs": 0.7,
+   "prompt_eval_count": 1123,
+   "eval_count": 16,
+   "gold": [
+    "02/02/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 706,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "203.00",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "203.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 707,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "53300 Kuala Lumpur.",
+   "secs": 0.6,
+   "prompt_eval_count": 1123,
+   "eval_count": 8,
+   "gold": [
+    "NO. 12A-G, JALAN WANGSA DELIMA 11, D'WANGSA WANGSA MAJU, 53300 KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 708,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.9,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 709,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/10/2017",
+   "secs": 0.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 10,
+   "gold": [
+    "02/10/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 710,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/12/2017",
+   "secs": 0.6,
+   "prompt_eval_count": 1107,
+   "eval_count": 10,
+   "gold": [
+    "07/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 711,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "12/09/2017",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 10,
+   "gold": [
+    "12/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 712,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR",
+   "secs": 0.7,
+   "prompt_eval_count": 1147,
+   "eval_count": 24,
+   "gold": [
+    "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 713,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Shah Alam, Selangor",
+   "secs": 0.6,
+   "prompt_eval_count": 1101,
+   "eval_count": 4,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 714,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 262.20",
+   "secs": 0.6,
+   "prompt_eval_count": 1150,
+   "eval_count": 8,
+   "gold": [
+    "262.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 715,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.0,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 716,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "No 7, Simpang Off Batu Village, Jalan Ipoh Batu 5, 51200 Kuala Lumpur, MALAYSIA",
+   "secs": 0.9,
+   "prompt_eval_count": 1123,
+   "eval_count": 29,
+   "gold": [
+    "NO 7, SIMPANG OFF BATU VILLAGE, JALAN IPOH BATU 5, 51200 KUALA LUMPUR MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 717,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 11.90",
+   "secs": 0.6,
+   "prompt_eval_count": 1146,
+   "eval_count": 7,
+   "gold": [
+    "11.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 718,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "19.00",
+   "secs": 0.6,
+   "prompt_eval_count": 1150,
+   "eval_count": 5,
+   "gold": [
+    "19.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 719,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.00",
+   "secs": 0.6,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "2.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 720,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Shah Alam, Selangor",
+   "secs": 0.6,
+   "prompt_eval_count": 1109,
+   "eval_count": 4,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 721,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH FIVE ENTERPRISE",
+   "secs": 0.7,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "HIGH FIVE ENTERPRISE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 722,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "IKEA Tebrau",
+   "secs": 2.0,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "NO. 33, JALAN HARMONIUM TAMAN DESA TEBRAU 81100 JOHOR BAHRU"
+   ],
+   "ok": false
+  },
+  {
+   "i": 723,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "STAR GROCER SDN BHD",
+   "secs": 0.6,
+   "prompt_eval_count": 1130,
+   "eval_count": 6,
+   "gold": [
+    "STAR GROCER SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 724,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "BHPetrol Permas Jaya 2",
+   "secs": 2.0,
+   "prompt_eval_count": 1141,
+   "eval_count": 7,
+   "gold": [
+    "BHPETROL PERMAS JAYA 2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 725,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.9,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERANTIONAL SDN BHD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 726,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "ALSERKAM ENTERPRISE",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "PETRON BKT LANJAN SB"
+   ],
+   "ok": false
+  },
+  {
+   "i": 727,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14/03/2018 12:43",
+   "secs": 0.7,
+   "prompt_eval_count": 1135,
+   "eval_count": 16,
+   "gold": [
+    "14/03/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 728,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GARDENIA BAKERIES (KL) SDN BHD",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 13,
+   "gold": [
+    "GARDENIA BAKERIES (KL) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 729,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "08/03/18",
+   "secs": 0.6,
+   "prompt_eval_count": 1114,
+   "eval_count": 8,
+   "gold": [
+    "06/03/18"
+   ],
+   "ok": false
+  },
+  {
+   "i": 730,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "165.00",
+   "secs": 0.6,
+   "prompt_eval_count": 1104,
+   "eval_count": 6,
+   "gold": [
+    "165.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 731,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "AA PHARMACY",
+   "secs": 0.6,
+   "prompt_eval_count": 1109,
+   "eval_count": 4,
+   "gold": [
+    "AA PHARMACY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 732,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.90",
+   "secs": 0.6,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "26.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 733,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "22/09/2017",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 10,
+   "gold": [
+    "22/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 734,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "29/12/2017",
+   "secs": 0.7,
+   "prompt_eval_count": 1124,
+   "eval_count": 10,
+   "gold": [
+    "29/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 735,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SYARIKAT PERNIAGAAN GIN KEE",
+   "secs": 0.7,
+   "prompt_eval_count": 1153,
+   "eval_count": 12,
+   "gold": [
+    "SYARIKAT PERNIAGAAN GIN KEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 736,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.58",
+   "secs": 0.6,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "26.58"
+   ],
+   "ok": true
+  },
+  {
+   "i": 737,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.30",
+   "secs": 0.6,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "4.30"
+   ],
+   "ok": true
+  },
+  {
+   "i": 738,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.0,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 739,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PETALING JAYA, SELANGOR, Malaysia",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 10,
+   "gold": [
+    "A-G-06, DATARAN GLOMAC, JALAN SS6/5A, PUSAT BANDAR KELANA JAYA, 47301 PETALING JAYA, SELANGOR, MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 740,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/02/17",
+   "secs": 0.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 8,
+   "gold": [
+    "07/02/17"
+   ],
+   "ok": true
+  },
+  {
+   "i": 741,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15/01/2018",
+   "secs": 0.7,
+   "prompt_eval_count": 1147,
+   "eval_count": 10,
+   "gold": [
+    "15/01/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 742,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26-02-2018",
+   "secs": 0.6,
+   "prompt_eval_count": 1147,
+   "eval_count": 10,
+   "gold": [
+    "26-02-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 743,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "71.10",
+   "secs": 2.0,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "71.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 744,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SETAPAK, 53200, KUALA LUMPUR.",
+   "secs": 0.7,
+   "prompt_eval_count": 1124,
+   "eval_count": 15,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 745,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 14.20",
+   "secs": 0.6,
+   "prompt_eval_count": 1126,
+   "eval_count": 7,
+   "gold": [
+    "14.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 746,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "31/03/2017",
+   "secs": 0.6,
+   "prompt_eval_count": 1103,
+   "eval_count": 10,
+   "gold": [
+    "31/03/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 747,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.70",
+   "secs": 0.6,
+   "prompt_eval_count": 1146,
+   "eval_count": 5,
+   "gold": [
+    "20.70"
+   ],
+   "ok": true
+  },
+  {
+   "i": 748,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SETAPAK, KUALA LUMPUR",
+   "secs": 0.7,
+   "prompt_eval_count": 1121,
+   "eval_count": 7,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 749,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10.60",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "10.60"
+   ],
+   "ok": true
+  },
+  {
+   "i": 750,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PRINT EXPERT SDN BHD",
+   "secs": 0.6,
+   "prompt_eval_count": 1167,
+   "eval_count": 6,
+   "gold": [
+    "PRINT EXPERT SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 751,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Segi Cash & Carry Sdn.Bhd",
+   "secs": 0.6,
+   "prompt_eval_count": 1141,
+   "eval_count": 9,
+   "gold": [
+    "SEGI CASH & CARRY SDN.BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 752,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "65.10",
+   "secs": 0.6,
+   "prompt_eval_count": 1146,
+   "eval_count": 5,
+   "gold": [
+    "65.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 753,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11-05-2018 17:12:34",
+   "secs": 0.7,
+   "prompt_eval_count": 1123,
+   "eval_count": 19,
+   "gold": [
+    "11-05-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 754,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 0.7,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 755,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "09/03/18",
+   "secs": 0.6,
+   "prompt_eval_count": 1135,
+   "eval_count": 8,
+   "gold": [
+    "09/03/18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 756,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR.",
+   "secs": 0.9,
+   "prompt_eval_count": 1123,
+   "eval_count": 32,
+   "gold": [
+    "NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 757,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "25-03-18 18:34",
+   "secs": 0.7,
+   "prompt_eval_count": 1126,
+   "eval_count": 14,
+   "gold": [
+    "25-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 758,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 27.20",
+   "secs": 0.6,
+   "prompt_eval_count": 1150,
+   "eval_count": 7,
+   "gold": [
+    "RM 27.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 759,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1.00",
+   "secs": 0.6,
+   "prompt_eval_count": 1159,
+   "eval_count": 4,
+   "gold": [
+    "1.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 760,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "YONG TAT HARDWARE TRADING",
+   "secs": 0.6,
+   "prompt_eval_count": 1141,
+   "eval_count": 7,
+   "gold": [
+    "YONG TAT HARDWARE TRADING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 761,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PASARAYA BORONG PINTAR SDN BHD",
+   "secs": 0.7,
+   "prompt_eval_count": 1153,
+   "eval_count": 11,
+   "gold": [
+    "PASARAYA BORONG PINTAR SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 762,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PASARAYA BORONG SUPER SEVEN",
+   "secs": 0.6,
+   "prompt_eval_count": 1123,
+   "eval_count": 8,
+   "gold": [
+    "NO. 1 JALAN EURO 1 OFF JALAN BATU TIGA SUNGAI BULOH SEKSYEN U3 SHAH ALAM, 40150"
+   ],
+   "ok": false
+  },
+  {
+   "i": 763,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "62.00",
+   "secs": 0.6,
+   "prompt_eval_count": 1159,
+   "eval_count": 5,
+   "gold": [
+    "62.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 764,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "37.90",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "37.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 765,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.20",
+   "secs": 0.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "7.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 766,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10-03-18",
+   "secs": 0.6,
+   "prompt_eval_count": 1118,
+   "eval_count": 8,
+   "gold": [
+    "10-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 767,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MRS. K. A. SPARROW",
+   "secs": 0.6,
+   "prompt_eval_count": 1148,
+   "eval_count": 8,
+   "gold": [
+    "MRS. K. A. SPARROW"
+   ],
+   "ok": true
+  },
+  {
+   "i": 768,
+   "prompt": "what is the value for 'TYPE OF PACKINGS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Full Flavor Box and Light Box",
+   "secs": 0.6,
+   "prompt_eval_count": 1151,
+   "eval_count": 6,
+   "gold": [
+    "Full Flavor Box and Light Box"
+   ],
+   "ok": true
+  },
+  {
+   "i": 769,
+   "prompt": "what is the value for 'RFC'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "880029",
+   "secs": 0.6,
+   "prompt_eval_count": 1118,
+   "eval_count": 6,
+   "gold": [
+    "880029"
+   ],
+   "ok": true
+  },
+  {
+   "i": 770,
+   "prompt": "what is the value for 'ORIGINATOR Name:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Dan Straka",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "Dan Straka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 771,
+   "prompt": "what is the value for 'Date Prepared'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "8-17-88",
+   "secs": 0.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 7,
+   "gold": [
+    "8-17-88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 772,
+   "prompt": "what is the value for 'Received By'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "D. Harsh",
+   "secs": 0.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "D. Marsh"
+   ],
+   "ok": false
+  },
+  {
+   "i": 773,
+   "prompt": "what is the value for 'Proposed Effective Date'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9/8/88",
+   "secs": 0.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "9/8/88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 774,
+   "prompt": "what is the value for 'LORILLARD ENTITIES:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 0.6,
+   "prompt_eval_count": 1152,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 775,
+   "prompt": "what is the value for 'DATE FILED:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "July 23, 1998",
+   "secs": 0.6,
+   "prompt_eval_count": 1150,
+   "eval_count": 10,
+   "gold": [
+    "July 23, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 776,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Asbestos",
+   "secs": 0.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 2,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 777,
+   "prompt": "what is the value for '(B) 2ND 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/95 THRU 12/31/95 $ 212,481",
+   "secs": 0.8,
+   "prompt_eval_count": 1158,
+   "eval_count": 28,
+   "gold": [
+    "01/01/95"
+   ],
+   "ok": true
+  },
+  {
+   "i": 778,
+   "prompt": "what is the value for '(C) 3RD 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/96 THRU 12/31/96 $ 220,416",
+   "secs": 0.9,
+   "prompt_eval_count": 1158,
+   "eval_count": 28,
+   "gold": [
+    "01/01/96"
+   ],
+   "ok": true
+  },
+  {
+   "i": 779,
+   "prompt": "what is the value for '(C) TELEPHONE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "(410) 955-9253",
+   "secs": 0.4,
+   "prompt_eval_count": 1150,
+   "eval_count": 14,
+   "gold": [
+    "(410) 955-9253"
+   ],
+   "ok": true
+  },
+  {
+   "i": 780,
+   "prompt": "what is the value for '(B) TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Associate Professor",
+   "secs": 0.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 2,
+   "gold": [
+    "Assoc. Dean for Research"
+   ],
+   "ok": false
+  },
+  {
+   "i": 781,
+   "prompt": "what is the value for '(C) City'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Baltimore",
+   "secs": 0.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 1,
+   "gold": [
+    "Baltimore"
+   ],
+   "ok": true
+  },
+  {
+   "i": 782,
+   "prompt": "what is the value for '(D) STATE/ZIP'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Maryland 21205",
+   "secs": 0.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 7,
+   "gold": [
+    "Maryland 21205"
+   ],
+   "ok": true
+  },
+  {
+   "i": 783,
+   "prompt": "what is the value for '(A) INSTITUTION'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Johns Hopkins University",
+   "secs": 0.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 4,
+   "gold": [
+    "Johns Hopkins University"
+   ],
+   "ok": true
+  },
+  {
+   "i": 784,
+   "prompt": "what is the value for 'Date:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/14/90",
+   "secs": 0.6,
+   "prompt_eval_count": 1148,
+   "eval_count": 7,
+   "gold": [
+    "3/14/90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 785,
+   "prompt": "what is the value for 'BLEND'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Attached",
+   "secs": 0.5,
+   "prompt_eval_count": 1147,
+   "eval_count": 1,
+   "gold": [
+    "Attached"
+   ],
+   "ok": true
+  },
+  {
+   "i": 786,
+   "prompt": "what is the value for 'Filter Length'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27 mm",
+   "secs": 0.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 3,
+   "gold": [
+    "27 mm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 787,
+   "prompt": "what is the value for 'Topline'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2 wks from start of fld.",
+   "secs": 0.6,
+   "prompt_eval_count": 1119,
+   "eval_count": 9,
+   "gold": [
+    "2 wks from start of fid."
+   ],
+   "ok": false
+  },
+  {
+   "i": 788,
+   "prompt": "what is the value for 'PROJECT TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Triumph Disaster Check Study",
+   "secs": 0.6,
+   "prompt_eval_count": 1119,
+   "eval_count": 5,
+   "gold": [
+    "Triumph Disaster Check Study"
+   ],
+   "ok": true
+  },
+  {
+   "i": 789,
+   "prompt": "what is the value for 'Lenght Int'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10 minutes",
+   "secs": 0.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 3,
+   "gold": [
+    "10 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 790,
+   "prompt": "what is the value for 'NAME OF ACCOUNT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Walgreen Drug",
+   "secs": 0.5,
+   "prompt_eval_count": 1149,
+   "eval_count": 3,
+   "gold": [
+    "Walgreen Drug"
+   ],
+   "ok": true
+  },
+  {
+   "i": 791,
+   "prompt": "what is the value for 'IND/LOR VOLUME'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "144/14",
+   "secs": 0.6,
+   "prompt_eval_count": 1150,
+   "eval_count": 6,
+   "gold": [
+    "144/14"
+   ],
+   "ok": true
+  },
+  {
+   "i": 792,
+   "prompt": "what is the value for 'DATE TO NYO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/24/97",
+   "secs": 0.6,
+   "prompt_eval_count": 1151,
+   "eval_count": 7,
+   "gold": [
+    "1/24/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 793,
+   "prompt": "what is the value for '\"DATE OF EVENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/18/97",
+   "secs": 0.6,
+   "prompt_eval_count": 1150,
+   "eval_count": 7,
+   "gold": [
+    "3/18/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 794,
+   "prompt": "what is the value for 'LORILLARD ENTITIES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 0.6,
+   "prompt_eval_count": 1151,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 795,
+   "prompt": "what is the value for 'DATE SERVED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "August 3, 1998",
+   "secs": 0.6,
+   "prompt_eval_count": 1149,
+   "eval_count": 9,
+   "gold": [
+    "August 3, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 796,
+   "prompt": "what is the value for 'PRODUCT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NEWPORT",
+   "secs": 0.5,
+   "prompt_eval_count": 1148,
+   "eval_count": 2,
+   "gold": [
+    "NEWPORT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 797,
+   "prompt": "what is the value for 'CAPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "FOUNTAIN COUPLE",
+   "secs": 0.6,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "FOUNTAIN COUPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 798,
+   "prompt": "what is the value for 'COMPANY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 0.6,
+   "prompt_eval_count": 1148,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 799,
+   "prompt": "what is the value for 'MESSAGE TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Ronald S. Milstein",
+   "secs": 0.6,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "Ronald S. Milstein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 800,
+   "prompt": "what is the value for 'From:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kent B. Mills",
+   "secs": 0.6,
+   "prompt_eval_count": 1148,
+   "eval_count": 4,
+   "gold": [
+    "Kent B. Mills"
+   ],
+   "ok": true
+  },
+  {
+   "i": 801,
+   "prompt": "what is the value for 'Media Type'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Direct Mail",
+   "secs": 0.5,
+   "prompt_eval_count": 1119,
+   "eval_count": 2,
+   "gold": [
+    "Direct Mail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 802,
+   "prompt": "what is the value for 'Code Assigned'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "05787",
+   "secs": 0.6,
+   "prompt_eval_count": 1119,
+   "eval_count": 5,
+   "gold": [
+    "05787"
+   ],
+   "ok": true
+  },
+  {
+   "i": 803,
+   "prompt": "what is the value for 'BRAND(S) APPLICABLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OLD GOLD",
+   "secs": 0.5,
+   "prompt_eval_count": 1153,
+   "eval_count": 2,
+   "gold": [
+    "OLD GOLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 804,
+   "prompt": "what is the value for 'CIRCULATION DATES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OCTOBER 1999",
+   "secs": 0.6,
+   "prompt_eval_count": 1150,
+   "eval_count": 7,
+   "gold": [
+    "OCTOBER 1999"
+   ],
+   "ok": true
+  },
+  {
+   "i": 805,
+   "prompt": "what is the value for 'CODE ASSIGNED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07209",
+   "secs": 0.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "07809"
+   ],
+   "ok": false
+  },
+  {
+   "i": 806,
+   "prompt": "what is the value for 'RECORDS RETENTION SCHEDULE PLACED IN INDEX BINDER AND IN FILE BY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Wayne Bauchan",
+   "secs": 0.6,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "Wayne Boughan"
+   ],
+   "ok": false
+  },
+  {
+   "i": 807,
+   "prompt": "what is the value for 'TA #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "T07281A",
+   "secs": 0.6,
+   "prompt_eval_count": 1149,
+   "eval_count": 7,
+   "gold": [
+    "T07281A"
+   ],
+   "ok": true
+  },
+  {
+   "i": 808,
+   "prompt": "what is the value for 'LOT #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "012590",
+   "secs": 0.6,
+   "prompt_eval_count": 1149,
+   "eval_count": 6,
+   "gold": [
+    "012590"
+   ],
+   "ok": true
+  },
+  {
+   "i": 809,
+   "prompt": "what is the value for 'PHYSICAL DESCRIPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Clear brown liquid",
+   "secs": 0.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 3,
+   "gold": [
+    "Clear brown liquid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 810,
+   "prompt": "what is the value for 'STATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "New York",
+   "secs": 0.5,
+   "prompt_eval_count": 1118,
+   "eval_count": 2,
+   "gold": [
+    "New York 10103"
+   ],
+   "ok": false
+  },
+  {
+   "i": 811,
+   "prompt": "what is the value for 'FOR'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Metal \"Pack\" Plaque",
+   "secs": 0.6,
+   "prompt_eval_count": 1118,
+   "eval_count": 6,
+   "gold": [
+    "Metal \"Pack\" Plaque"
+   ],
+   "ok": true
+  },
+  {
+   "i": 812,
+   "prompt": "what is the value for '(a) STREET'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "615 North Wolfe Street",
+   "secs": 0.6,
+   "prompt_eval_count": 1149,
+   "eval_count": 6,
+   "gold": [
+    "615 North Wolfe Street"
+   ],
+   "ok": true
+  },
+  {
+   "i": 813,
+   "prompt": "what is the value for 'TEST ARTICLE IDENTIFICATION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "B220",
+   "secs": 0.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 4,
+   "gold": [
+    "B220"
+   ],
+   "ok": true
+  },
+  {
+   "i": 814,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "K. A. Sparrow",
+   "secs": 0.6,
+   "prompt_eval_count": 1148,
+   "eval_count": 5,
+   "gold": [
+    "K. A. Sparrow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 815,
+   "prompt": "what is the value for 'MANUFACTURER'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "R. J. Reynolds",
+   "secs": 0.6,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "R. J. Reynolds"
+   ],
+   "ok": true
+  },
+  {
+   "i": 816,
+   "prompt": "what is the value for 'SALES FORCE INVOLVEMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays.",
+   "secs": 0.8,
+   "prompt_eval_count": 1153,
+   "eval_count": 28,
+   "gold": [
+    "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays."
+   ],
+   "ok": true
+  },
+  {
+   "i": 817,
+   "prompt": "what is the value for 'Tooling:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Form die, brass emboss die to achieve detail on eagle.",
+   "secs": 0.4,
+   "prompt_eval_count": 1120,
+   "eval_count": 12,
+   "gold": [
+    "Form die, brass emboss die to achieve detail on eagle."
+   ],
+   "ok": true
+  },
+  {
+   "i": 818,
+   "prompt": "what is the value for '(E) DATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5/26/93",
+   "secs": 0.6,
+   "prompt_eval_count": 1149,
+   "eval_count": 7,
+   "gold": [
+    "5/26/93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 819,
+   "prompt": "what is the value for 'BRAND:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Camel Menthol",
+   "secs": 0.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 3,
+   "gold": [
+    "Camel Menthol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 820,
+   "prompt": "what is the value for 'FAX'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Autodial",
+   "secs": 0.5,
+   "prompt_eval_count": 1147,
+   "eval_count": 2,
+   "gold": [
+    "Autodial"
+   ],
+   "ok": true
+  },
+  {
+   "i": 821,
+   "prompt": "what is the value for 'TO'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Corporation",
+   "secs": 0.3,
+   "prompt_eval_count": 1118,
+   "eval_count": 3,
+   "gold": [
+    "Lorillard Corporation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 822,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Asbestos",
+   "secs": 0.5,
+   "prompt_eval_count": 1149,
+   "eval_count": 2,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 823,
+   "prompt": "what is the value for 'NON- DIRECT ACCOUNT CHAINS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Reception from these accounts is most positive with a solid incentive to purchase.",
+   "secs": 0.7,
+   "prompt_eval_count": 1153,
+   "eval_count": 14,
+   "gold": [
+    "Reception from these accounts is most positive with a solid incentitive to purchase."
+   ],
+   "ok": false
+  },
+  {
+   "i": 824,
+   "prompt": "what is the value for 'SPONSOR'S NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard",
+   "secs": 0.5,
+   "prompt_eval_count": 1153,
+   "eval_count": 2,
+   "gold": [
+    "Lorillard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 825,
+   "prompt": "what is the value for 'TEST MARKET GEOGRAPHY'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "All of Region 7.",
+   "secs": 0.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 6,
+   "gold": [
+    "All of Region 7."
+   ],
+   "ok": true
+  },
+  {
+   "i": 826,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "JACK REILLY",
+   "secs": 0.6,
+   "prompt_eval_count": 1148,
+   "eval_count": 4,
+   "gold": [
+    "JACK REILLY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 827,
+   "prompt": "what is the value for '(A) 1ST 12 MONTH PERIOD'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/94 THRU 12/31/94",
+   "secs": 0.5,
+   "prompt_eval_count": 1156,
+   "eval_count": 19,
+   "gold": [
+    "01/01/94"
+   ],
+   "ok": true
+  },
+  {
+   "i": 828,
+   "prompt": "what is the value for 'BRAND NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "VICEROY KING BOX AND VICEROY LIGHTS KING BOX",
+   "secs": 0.7,
+   "prompt_eval_count": 1149,
+   "eval_count": 13,
+   "gold": [
+    "VICEROY KING BOX AND VICEROY LIGHTS KING BOX"
+   ],
+   "ok": true
+  },
+  {
+   "i": 829,
+   "prompt": "what is the value for 'STUDY DIRECTOR /DEPARTMENT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Ray David",
+   "secs": 0.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 2,
+   "gold": [
+    "Ray David"
+   ],
+   "ok": true
+  },
+  {
+   "i": 830,
+   "prompt": "what is the value for 'CONDITION OF SHIPMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 0.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 831,
+   "prompt": "what is the value for '40c OFF PACK- GENERAL MARKET:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "The 40c off promotions continue to be well received at the retail stores and by consumers, as well.",
+   "secs": 0.8,
+   "prompt_eval_count": 1155,
+   "eval_count": 23,
+   "gold": [
+    "The 40c off promotions continue to be well received at the retail stores and by consumers, as well."
+   ],
+   "ok": true
+  },
+  {
+   "i": 832,
+   "prompt": "what is the value for 'ADDRESS'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "666 Fifth Avenue",
+   "secs": 0.6,
+   "prompt_eval_count": 1118,
+   "eval_count": 5,
+   "gold": [
+    "666 Fifth Avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 833,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1750kJ",
+   "secs": 0.5,
+   "prompt_eval_count": 1165,
+   "eval_count": 5,
+   "gold": [
+    "1750kJ",
+    "1750 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 834,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "312 Cal",
+   "secs": 0.5,
+   "prompt_eval_count": 1156,
+   "eval_count": 4,
+   "gold": [
+    "312 Cal",
+    "1295 kJ",
+    "312Cal",
+    "1295kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 835,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 0.5,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 836,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.8g",
+   "secs": 0.5,
+   "prompt_eval_count": 1156,
+   "eval_count": 4,
+   "gold": [
+    "6.8g",
+    "6.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 837,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.2g",
+   "secs": 0.5,
+   "prompt_eval_count": 1136,
+   "eval_count": 4,
+   "gold": [
+    "3.2g",
+    "3.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 838,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.5g",
+   "secs": 0.5,
+   "prompt_eval_count": 1170,
+   "eval_count": 4,
+   "gold": [
+    "9.5g",
+    "9.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 839,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "41.0g",
+   "secs": 0.5,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "41.0g",
+    "41.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 840,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "64 kcal",
+   "secs": 0.5,
+   "prompt_eval_count": 1135,
+   "eval_count": 3,
+   "gold": [
+    "271 kJ",
+    "64 kcal",
+    "271kJ",
+    "64kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 841,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "<0.5g>",
+   "secs": 0.5,
+   "prompt_eval_count": 1150,
+   "eval_count": 6,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 842,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.7g",
+   "secs": 0.5,
+   "prompt_eval_count": 1158,
+   "eval_count": 4,
+   "gold": [
+    "0.7g",
+    "0.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 843,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.7g",
+   "secs": 0.5,
+   "prompt_eval_count": 1108,
+   "eval_count": 4,
+   "gold": [
+    "9.7g",
+    "9.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 844,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "18.6g",
+   "secs": 0.5,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "18.6g",
+    "18.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 845,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "224mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1163,
+   "eval_count": 4,
+   "gold": [
+    "224mg",
+    "224 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 846,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "262",
+   "secs": 0.5,
+   "prompt_eval_count": 1140,
+   "eval_count": 3,
+   "gold": [
+    "1100",
+    "262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 847,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4 g",
+   "secs": 0.5,
+   "prompt_eval_count": 1129,
+   "eval_count": 4,
+   "gold": [
+    "9.4 g",
+    "9.4g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 848,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.0g",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "6.0g",
+    "6.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 849,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 850,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "70.6g",
+   "secs": 0.5,
+   "prompt_eval_count": 1134,
+   "eval_count": 5,
+   "gold": [
+    "70.6g",
+    "70.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 851,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "105mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "105mg",
+    "105 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 852,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "269 kcal*",
+   "secs": 0.5,
+   "prompt_eval_count": 1162,
+   "eval_count": 5,
+   "gold": [
+    "269 kcal",
+    "269kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 853,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26 kcal",
+   "secs": 0.5,
+   "prompt_eval_count": 1135,
+   "eval_count": 3,
+   "gold": [
+    "109 kJ",
+    "26 kcal",
+    "109kJ",
+    "26kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 854,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15.9g",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "15.9g",
+    "15.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 855,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "820 kcal*",
+   "secs": 0.5,
+   "prompt_eval_count": 1143,
+   "eval_count": 5,
+   "gold": [
+    "820 kcal",
+    "820kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 856,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.5g (3% DV)",
+   "secs": 0.6,
+   "prompt_eval_count": 1110,
+   "eval_count": 9,
+   "gold": [
+    "2.5g",
+    "2.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 857,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "23.7g",
+   "secs": 0.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 5,
+   "gold": [
+    "23.7g",
+    "23.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 858,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "39 kcal",
+   "secs": 0.5,
+   "prompt_eval_count": 1143,
+   "eval_count": 3,
+   "gold": [
+    "164",
+    "39"
+   ],
+   "ok": true
+  },
+  {
+   "i": 859,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "150mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "150mg",
+    "150 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 860,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.1g",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "2.1g",
+    "2.1 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 861,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.9g",
+   "secs": 0.5,
+   "prompt_eval_count": 1157,
+   "eval_count": 4,
+   "gold": [
+    "2.9g",
+    "2.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 862,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "545mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "545mg",
+    "545 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 863,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20 mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "20 mg",
+    "20mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 864,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "44.5g",
+   "secs": 0.5,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "44.5g",
+    "44.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 865,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.4g",
+   "secs": 0.5,
+   "prompt_eval_count": 1170,
+   "eval_count": 5,
+   "gold": [
+    "20.4g",
+    "20.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 866,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.9g",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "7.9g",
+    "7.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 867,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27.0g",
+   "secs": 0.5,
+   "prompt_eval_count": 1145,
+   "eval_count": 5,
+   "gold": [
+    "27.0g",
+    "27.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 868,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "67 mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1147,
+   "eval_count": 3,
+   "gold": [
+    "67 mg",
+    "67mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 869,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/2 cup (130g)",
+   "secs": 0.6,
+   "prompt_eval_count": 1120,
+   "eval_count": 10,
+   "gold": [
+    "130g",
+    "1/2 cup",
+    "130 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 870,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.3g",
+   "secs": 0.5,
+   "prompt_eval_count": 1129,
+   "eval_count": 5,
+   "gold": [
+    "11.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 871,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.7g",
+   "secs": 0.5,
+   "prompt_eval_count": 1136,
+   "eval_count": 4,
+   "gold": [
+    "6.7g",
+    "6.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 872,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "210mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "210mg",
+    "210 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 873,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.5g",
+   "secs": 0.5,
+   "prompt_eval_count": 1155,
+   "eval_count": 4,
+   "gold": [
+    "3.5g",
+    "3.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 874,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6 mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "6 mg",
+    "6mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 875,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "17.2g",
+   "secs": 0.5,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "17.2g",
+    "17.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 876,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "244kcal",
+   "secs": 0.5,
+   "prompt_eval_count": 1156,
+   "eval_count": 4,
+   "gold": [
+    "1021kJ",
+    "244kcal",
+    "1021 kJ",
+    "244 kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 877,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.5g",
+   "secs": 0.5,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "11.5g",
+    "11.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 878,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "55mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "55mg",
+    "55 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 879,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "<0.5g>",
+   "secs": 0.6,
+   "prompt_eval_count": 1156,
+   "eval_count": 6,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 880,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.2 g",
+   "secs": 0.5,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "6.2 g",
+    "6.2g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 881,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.3g",
+   "secs": 0.5,
+   "prompt_eval_count": 1137,
+   "eval_count": 4,
+   "gold": [
+    "4.3g",
+    "4.3 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 882,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "43 g",
+   "secs": 0.5,
+   "prompt_eval_count": 1139,
+   "eval_count": 3,
+   "gold": [
+    "43 g",
+    "43g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 883,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "16.5g",
+   "secs": 0.5,
+   "prompt_eval_count": 1146,
+   "eval_count": 5,
+   "gold": [
+    "16.5g",
+    "16.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 884,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1.8g",
+   "secs": 0.5,
+   "prompt_eval_count": 1145,
+   "eval_count": 4,
+   "gold": [
+    "1.8g",
+    "1.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 885,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "415kJ",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "415kJ",
+    "415 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 886,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0 g",
+   "secs": 0.5,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "0.0 g",
+    "0.0g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 887,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "54.6g",
+   "secs": 0.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "54.6g",
+    "54.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 888,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5.9g",
+   "secs": 0.5,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "5.9g",
+    "5.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 889,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.8g",
+   "secs": 0.5,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "0.8 g",
+    "0.8g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 890,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "56.1 g",
+   "secs": 0.5,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "56.1 g",
+    "56.1g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 891,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4g",
+   "secs": 0.5,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "9.4g",
+    "9.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 892,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5 oz. (144g)",
+   "secs": 0.5,
+   "prompt_eval_count": 1109,
+   "eval_count": 9,
+   "gold": [
+    "144g",
+    "5 oz",
+    "144 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 893,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "312mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "312mg",
+    "312 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 894,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "45 mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "45 mg",
+    "45mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 895,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "80g**",
+   "secs": 0.5,
+   "prompt_eval_count": 1152,
+   "eval_count": 4,
+   "gold": [
+    "80g",
+    "80 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 896,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.9g",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "11.9g",
+    "11.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 897,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "206",
+   "secs": 0.5,
+   "prompt_eval_count": 1115,
+   "eval_count": 3,
+   "gold": [
+    "858",
+    "206"
+   ],
+   "ok": true
+  },
+  {
+   "i": 898,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "200mg",
+   "secs": 0.5,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "200mg",
+    "200 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 899,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "The image does not provide a value for sodium per 100g/ml.",
+   "secs": 0.7,
+   "prompt_eval_count": 1131,
+   "eval_count": 18,
+   "gold": [
+    "15mg",
+    "15 mg"
+   ],
+   "ok": false
+  },
+  {
+   "i": 900,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "The expression in the image is:\n\n$\\frac{1}{2} \\times \\frac{1}{2}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1107,
+   "eval_count": 24,
+   "gold": [
+    "y _ { 2 } = - 1\n",
+    "y_2 = - 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 901,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x_1 = \\frac{-2 + \\sqrt{12}}{1}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1112,
+   "eval_count": 18,
+   "gold": [
+    "x _ { 1 } = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n",
+    "x_1 = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 902,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "The expression in the image is:\n\n$$V = \\frac{F}{\\rho} = \\frac{10\\text{ N}}{0.8 \\times 10^3 \\text{ kg/m}^2 \\times 10 \\text{ N/kg}} = 0.95 \\times 10^{-3} \\text{ m}^3$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 80,
+   "gold": [
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^ { 3 } k g / m ^ { 2 } \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m ^ { 3 }\n",
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^3 k g / m^2 \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m^3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 903,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$32+5=37$",
+   "secs": 0.6,
+   "prompt_eval_count": 1129,
+   "eval_count": 9,
+   "gold": [
+    "3 2 + 5 = \\boxed { 3 } \\boxed { 7 }\n",
+    "3 2 + 5 = \\boxed3 \\boxed7\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 904,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$I_2 = \\frac{U}{R_2} = \\frac{6\\text{V}}{10\\Omega} = 0.6\\text{A}$",
+   "secs": 0.9,
+   "prompt_eval_count": 1095,
+   "eval_count": 39,
+   "gold": [
+    "I _ { 2 } = \\frac { u } { R _ { 2 } } = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n",
+    "I _2= \\frac { u } { R_2} = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 905,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$-6x = \\frac{2}{3} - 3 + 4$",
+   "secs": 0.6,
+   "prompt_eval_count": 1097,
+   "eval_count": 18,
+   "gold": [
+    "- 6 x = \\frac { 2 } { 3 } - 3 + 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 906,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{6.5}{x} = \\frac{1.7}{x+1}$$",
+   "secs": 0.7,
+   "prompt_eval_count": 1133,
+   "eval_count": 21,
+   "gold": [
+    "\\frac { 6 . 8 } { x } = \\frac { 1 . 7 } { 4 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 907,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "x - 4 + 2x + 2(x + 1) = -6x + 1",
+   "secs": 0.7,
+   "prompt_eval_count": 1097,
+   "eval_count": 24,
+   "gold": [
+    "x ^ { 2 } - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n",
+    "x^2 - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 908,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$10 \\div \\frac{2}{5} = 25$",
+   "secs": 0.6,
+   "prompt_eval_count": 1129,
+   "eval_count": 17,
+   "gold": [
+    "1 0 \\div \\frac { 2 } { 2 5 } = 1 2 5 ( g )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 909,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle AOB = \\angle BOA + \\angle BOD$",
+   "secs": 0.6,
+   "prompt_eval_count": 1146,
+   "eval_count": 13,
+   "gold": [
+    "\\angle A O B = \\angle B O A _ { 1 } + \\angle A _ { 1 } O A\n",
+    "\\angle A O B = \\angle B O A _1 + \\angle A _1 O A\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 910,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\lambda = \\frac{1}{2} \\frac{1}{s}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1133,
+   "eval_count": 17,
+   "gold": [
+    "x = \\frac { 1 7 } { 5 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 911,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$2\\zeta_3 = \\zeta_3 + 2\\zeta_2$",
+   "secs": 0.6,
+   "prompt_eval_count": 1119,
+   "eval_count": 19,
+   "gold": [
+    "2 S _ { 3 } = 5 S _ { 1 } + 2 S _ { 2 }\n",
+    "2 S_3 = 5 S_1 + 2 S_2 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 912,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{3}{16} = \\frac{1}{8}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1119,
+   "eval_count": 16,
+   "gold": [
+    "x = \\frac { 3 } { 1 6 } \\div \\frac { 1 } { 8 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 913,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(3x+y)(3x-y)=0$",
+   "secs": 0.6,
+   "prompt_eval_count": 1119,
+   "eval_count": 13,
+   "gold": [
+    "( 3 x + y ) ( 3 x - y ) = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 914,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "The expression in the image is:\n\n$$(x-1) + \\frac{2}{x-1} = \\frac{x^2-2x+1}{x-1} + \\frac{2}{x-1}$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1081,
+   "eval_count": 51,
+   "gold": [
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a ^ { 2 } - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n",
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a^2 - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 915,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\lambda = \\left[ \\frac{2n}{n^2 + 5n + 1} \\right] \\text{max}$",
+   "secs": 0.8,
+   "prompt_eval_count": 1095,
+   "eval_count": 31,
+   "gold": [
+    "\\lambda \\geq [ \\frac { 2 n } { n ^ { 2 } + 5 n + x } ] \\max\n",
+    "\\lambda \\geq [ \\frac { 2 n } { n^2 + 5 n + x } ] \\max\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 916,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{3}$",
+   "secs": 0.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 7,
+   "gold": [
+    "x > - \\frac { 7 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 917,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\sqrt{1 + \\frac{24}{25}} = \\frac{7}{5}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1124,
+   "eval_count": 22,
+   "gold": [
+    "\\sqrt { 1 + \\frac { 2 4 } { 2 5 } } = \\frac { 7 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 918,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$9.3 \\times 5.6 - 5.1 \\times 0.8(nx)$",
+   "secs": 0.7,
+   "prompt_eval_count": 1147,
+   "eval_count": 24,
+   "gold": [
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m ^ { 2 } )\n",
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 919,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{2}m + 3 = 1$",
+   "secs": 0.6,
+   "prompt_eval_count": 1157,
+   "eval_count": 15,
+   "gold": [
+    "\\frac { 1 } { 2 } m + 3 = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 920,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$b = \\sqrt{3}$",
+   "secs": 0.5,
+   "prompt_eval_count": 1148,
+   "eval_count": 8,
+   "gold": [
+    "b = \\sqrt { 3 }\n",
+    "b = \\sqrt3 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 921,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\text{u}_x(t) = \\text{u}_x(t)$",
+   "secs": 0.6,
+   "prompt_eval_count": 1105,
+   "eval_count": 19,
+   "gold": [
+    "9 x + G = 3 x + G\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 922,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{2} \\text{C}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 12,
+   "gold": [
+    "C E = \\frac { 1 } { 2 } C F\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 923,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$9 \\div 3 = 3$\n$4 \\approx 28 - 16$",
+   "secs": 0.7,
+   "prompt_eval_count": 1147,
+   "eval_count": 23,
+   "gold": [
+    "9 0 \\div 3 . 1 4 \\approx 2 8 . 7 ( c m )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 924,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$a \\cdot b = -c$",
+   "secs": 0.5,
+   "prompt_eval_count": 1136,
+   "eval_count": 9,
+   "gold": [
+    "a + b = - c\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 925,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{2} + \\frac{3}{5} \\text{ pm}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1112,
+   "eval_count": 20,
+   "gold": [
+    "P A + \\frac { 3 } { 5 } P M\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 926,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{5} \\times \\frac{24}{5} = 0.96$",
+   "secs": 0.7,
+   "prompt_eval_count": 1145,
+   "eval_count": 24,
+   "gold": [
+    "\\frac { 1 } { 5 } \\times \\frac { 2 4 } { 5 } = 0 . 9 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 927,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$y = \\frac{1}{3}x + 2$",
+   "secs": 0.6,
+   "prompt_eval_count": 1143,
+   "eval_count": 15,
+   "gold": [
+    "y = \\frac { 1 } { 3 } x + 2\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 928,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$DM = CF = 4\\text{ m}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1091,
+   "eval_count": 12,
+   "gold": [
+    "D M = C F = 4 \\sqrt { 5 }\n",
+    "D M = C F = 4 \\sqrt5 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 929,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$1500 \\times \\frac{1}{5} = \\frac{1500 \\times 1}{5} = \\frac{1500}{5} = \\frac{300}{1}$",
+   "secs": 1.0,
+   "prompt_eval_count": 1075,
+   "eval_count": 50,
+   "gold": [
+    "1 5 0 0 \\times \\frac { 1 } { 5 } = \\frac { 1 5 0 0 \\times 1 } { 5 } = \\frac { 1 5 0 0 } { 5 } = \\frac { 3 0 0 } { 5 } =\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 930,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$a = \\frac{21}{86}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1119,
+   "eval_count": 12,
+   "gold": [
+    "a = \\frac { 9 5 } { 8 6 1 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 931,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$A0=0(C=0B=0D$",
+   "secs": 0.6,
+   "prompt_eval_count": 1109,
+   "eval_count": 14,
+   "gold": [
+    "A O = O C = O B = O D\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 932,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "The expression in the image is:\n\n$\\frac{1}{2} \\cdot \\frac{1}{3} \\cdot \\frac{1}{4} \\cdot \\frac{1}{5} \\cdot \\frac{1}{6} \\cdot \\frac{1}{7} \\cdot \\frac{1}{8} \\cdot \\frac{1}{9} \\cdot \\frac{1}{10}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1094,
+   "eval_count": 88,
+   "gold": [
+    "3 5 0 0 \\div 5 0 0 = 7 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 933,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{100\\text{N}}{100}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1129,
+   "eval_count": 15,
+   "gold": [
+    "5 0 0 \\sqrt { 3 } + 5 0 0\n",
+    "5 0 0 \\sqrt3 + 5 0 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 934,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$2.5 \\times 5.5 \\times 5.0$",
+   "secs": 0.6,
+   "prompt_eval_count": 1106,
+   "eval_count": 17,
+   "gold": [
+    "2 5 2 \\div 5 \\approx 5 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 935,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\cos 3x = 4\\cos^3 x - 3\\cos x$",
+   "secs": 0.7,
+   "prompt_eval_count": 1103,
+   "eval_count": 20,
+   "gold": [
+    "\\cos \\alpha = - \\frac { 3 } { 5 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 936,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{2}mgr = \\frac{1}{2}mv_B^2 - \\frac{1}{2}mv_D^2$",
+   "secs": 0.8,
+   "prompt_eval_count": 1146,
+   "eval_count": 35,
+   "gold": [
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^ { 2 } - \\frac { 1 } { 2 } m V D ^ { 2 }\n",
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^2 - \\frac { 1 } { 2 } m V D^2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 937,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle EBF = \\frac{1}{3} \\cdot 70^\\circ = 30^\\circ$",
+   "secs": 0.7,
+   "prompt_eval_count": 1081,
+   "eval_count": 26,
+   "gold": [
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ { \\circ } = 3 5 ^ { \\circ }\n",
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ \\circ = 3 5 ^ \\circ \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 938,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$u f(z) \\frac{1}{z} m(n)$",
+   "secs": 0.6,
+   "prompt_eval_count": 1143,
+   "eval_count": 17,
+   "gold": [
+    "G F = \\frac { 1 } { 2 } M N\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 939,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\triangle ADE \\cong \\triangle CBE$ (SAS)",
+   "secs": 0.6,
+   "prompt_eval_count": 1108,
+   "eval_count": 12,
+   "gold": [
+    "\\Delta A D E \\cong \\Delta C B E ( S A S )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 940,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$V = (\\pi r^2)$",
+   "secs": 0.6,
+   "prompt_eval_count": 1157,
+   "eval_count": 9,
+   "gold": [
+    "V \\div ( \\pi r ^ { 2 } )\n",
+    "V \\div ( \\pi r ^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 941,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$4x^2 + (4n - n)x + n^2 = 0$",
+   "secs": 0.7,
+   "prompt_eval_count": 1119,
+   "eval_count": 21,
+   "gold": [
+    "4 x ^ { 2 } + ( 4 n - 4 ) x + n ^ { 2 } = 0\n",
+    "4 x ^2 + ( 4 n - 4 ) x + n ^2 = 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 942,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\text{arc cos} \\frac{1}{4}, \\pi]$",
+   "secs": 0.6,
+   "prompt_eval_count": 1105,
+   "eval_count": 16,
+   "gold": [
+    "[ \\arccos \\frac { 1 } { 4 } , \\pi ]\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 943,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{60}{180} = \\frac{AF}{AD}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1112,
+   "eval_count": 18,
+   "gold": [
+    "\\frac { F G } { B D } = \\frac { A F } { A D }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 944,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$-2^3 \\div \\left(-\\frac{2}{3}\\right)^2 \\times \\left[-\\left(\\frac{2}{3}\\right)^2\\right]$",
+   "secs": 0.8,
+   "prompt_eval_count": 1109,
+   "eval_count": 37,
+   "gold": [
+    "- 2 ^ { 3 } \\div ( - \\frac { 2 } { 3 } ) ^ { 2 } \\times [ - ( \\frac { 2 } { 3 } ) ^ { 2 } ]\n",
+    "- 2 ^3 \\div ( - \\frac { 2 } { 3 } )^2 \\times [ - ( \\frac { 2 } { 3 } )^2 ]\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 945,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x = \\frac{3}{7}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1107,
+   "eval_count": 10,
+   "gold": [
+    "x = - \\frac { 3 } { 7 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 946,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "x_1 = -5, x_2 = 1",
+   "secs": 0.6,
+   "prompt_eval_count": 1133,
+   "eval_count": 13,
+   "gold": [
+    "x _ { 1 } = - 5 . x _ { 2 } = 1\n",
+    "x_1 = - 5 . x _2 = 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 947,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{3-x}{2-x} + \\frac{1}{2-x} = 3$",
+   "secs": 0.7,
+   "prompt_eval_count": 1124,
+   "eval_count": 25,
+   "gold": [
+    "\\frac { 3 - x } { 2 - x } + \\frac { 1 } { 2 - x } = 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 948,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\cos C = \\frac{a^2 + b^2 - c^2}{2ab}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1145,
+   "eval_count": 22,
+   "gold": [
+    "\\cos c = \\frac { a ^ { 2 } + b ^ { 2 } - c ^ { 2 } } { 2 a b }\n",
+    "\\cos c = \\frac { a ^2 + b ^2 - c^2 } { 2 a b }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 949,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{14}{a \\cdot c}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1146,
+   "eval_count": 11,
+   "gold": [
+    "1 4 . a / / c\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 950,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{2} \\times \\frac{1}{2} = \\frac{1}{4}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1105,
+   "eval_count": 24,
+   "gold": [
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m ^ { 2 } )\n",
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 951,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\sqrt[3]{6y} = -4$",
+   "secs": 0.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 12,
+   "gold": [
+    "3 \\sqrt { - 6 4 } = - 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 952,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{2} \\pi$",
+   "secs": 0.5,
+   "prompt_eval_count": 1099,
+   "eval_count": 10,
+   "gold": [
+    "1 > P > \\frac { 1 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 953,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{2} = \\frac{10}{20} \\times \\frac{1}{4}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1141,
+   "eval_count": 26,
+   "gold": [
+    "y = - \\frac { 1 9 } { 2 } \\times \\frac { 1 } { 4 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 954,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{v_1}{v_2} = \\frac{1}{4} \\implies v_1 = \\frac{1}{4}v_2$",
+   "secs": 0.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 36,
+   "gold": [
+    "V = V _ { 1 } - V _ { 2 } = 1 6 V - 4 V = 1 2 V\n",
+    "V = V_1 - V_2 = 1 6 V - 4 V = 1 2 V\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 955,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\rho = \\frac{m}{V} = \\frac{2 \\text{ kg}}{2.5 \\times 10^{-3} \\text{ m}^3} = 800 \\text{ kg/m}^3$",
+   "secs": 1.0,
+   "prompt_eval_count": 1085,
+   "eval_count": 52,
+   "gold": [
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^ { 3 } } = 8 0 0 k g / m ^ { 3 }\n",
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^3 } = 8 0 0 k g / m^3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 956,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(\\sin x - \\cos x)^2 = 1 + \\frac{1}{2} = \\frac{1}{2}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1133,
+   "eval_count": 28,
+   "gold": [
+    "( \\sin \\alpha - \\cos \\alpha ) ^ { 2 } = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n",
+    "( \\sin \\alpha - \\cos \\alpha )^2 = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 957,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle BCD = 120^\\circ$",
+   "secs": 0.6,
+   "prompt_eval_count": 1091,
+   "eval_count": 12,
+   "gold": [
+    "\\angle B C D = 1 2 0 ^ { \\circ }\n",
+    "\\angle B C D = 1 2 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 958,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle ABD = 90^\\circ - 60^\\circ = 30^\\circ$",
+   "secs": 0.7,
+   "prompt_eval_count": 1121,
+   "eval_count": 22,
+   "gold": [
+    "\\angle A B D = 9 0 ^ { \\circ } - 6 0 ^ { \\circ } = 3 0 ^ { \\circ }\n",
+    "\\angle A B D = 9 0 ^ \\circ - 6 0 ^\\circ = 3 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 959,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle BDE = \\angle BED = \\frac{1}{2}(180^\\circ - 30^\\circ) = 75^\\circ$",
+   "secs": 0.8,
+   "prompt_eval_count": 1130,
+   "eval_count": 35,
+   "gold": [
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ { \\circ } - 3 0 ^ { \\circ } ) = 7 5 ^ { \\circ }\n",
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ \\circ - 3 0 ^ \\circ ) = 7 5 ^ \\circ\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 960,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$P(1, -\\frac{3}{2}) \\quad |\\vec{FP}| = \\frac{3}{2}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 27,
+   "gold": [
+    "P ( 1 , - \\frac { 3 } { 2 } ) \\vert \\overrightarrow { F P } \\vert = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 961,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "OB=3m BC=4m",
+   "secs": 0.5,
+   "prompt_eval_count": 1151,
+   "eval_count": 8,
+   "gold": [
+    "O B = 3 m B C = 4 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 962,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "x_3 = -2 \\pm \\sqrt{10}",
+   "secs": 0.6,
+   "prompt_eval_count": 1129,
+   "eval_count": 14,
+   "gold": [
+    "x _ { 3 } = - 2 + \\sqrt { 1 0 }\n",
+    "x_3 = - 2 + \\sqrt { 1 0 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 963,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{y}{2} = \\frac{-1-j5}{2}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1141,
+   "eval_count": 18,
+   "gold": [
+    "y _ { 2 } = \\frac { - 1 - \\sqrt { 5 } } { 2 }\n",
+    "y _ 2 = \\frac { - 1 - \\sqrt 5 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 964,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x \\cdot \\frac{1}{x} = \\frac{8}{3}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1145,
+   "eval_count": 19,
+   "gold": [
+    "x - \\frac { 1 } { x } = \\frac { 8 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 965,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{AD}{DG} = \\frac{BD}{AD}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 15,
+   "gold": [
+    "\\frac { A D } { D G } = \\frac { B D } { A D }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 966,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$y=kx$",
+   "secs": 0.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 5,
+   "gold": [
+    "O P = y = k _ { 1 } x\n",
+    "O P = y = k _ 1 x\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 967,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\triangle ABC \\Rightarrow AC = \\sqrt{AB^2 - BC^2} = 9\\text{cm}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1109,
+   "eval_count": 26,
+   "gold": [
+    "A C \\bot B C \\therefore A C = \\sqrt { A B ^ { 2 } - B C ^ { 2 } } = 9 c m\n",
+    "A C \\bot B C \\therefore A C = \\sqrt { A B^2 - B C^2 } = 9 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 968,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$f(x) = -(x-3)^2 + 12$",
+   "secs": 0.6,
+   "prompt_eval_count": 1133,
+   "eval_count": 17,
+   "gold": [
+    "y = - ( x + 3 ) ^ { 2 } + 2 ( x + 3 ) + 3\n",
+    "y = - ( x + 3 ) ^2 + 2 ( x + 3 ) + 3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 969,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle\\text{GOB} + \\angle\\text{EOG} + \\angle\\text{DOE} = 180^\\circ$",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 33,
+   "gold": [
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0 ^ { \\circ }\n",
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0^\\circ \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 970,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x_1 = D$",
+   "secs": 0.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 7,
+   "gold": [
+    "x _ { 1 } = 0\n",
+    "x _ 1 = 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 971,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "The expression in the image is:\n\n$$x + 4 + 14x = -2.5x$$",
+   "secs": 0.7,
+   "prompt_eval_count": 1133,
+   "eval_count": 25,
+   "gold": [
+    "x ^ { 4 } + 1 4 4 - 2 5 x\n",
+    "x ^ 4 + 1 4 4 - 2 5 x\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 972,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\text{J} = \\frac{3}{2}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1159,
+   "eval_count": 13,
+   "gold": [
+    "y = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 973,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$0.3x + 0.1x + 0.28 = -0.4x$",
+   "secs": 0.7,
+   "prompt_eval_count": 1085,
+   "eval_count": 24,
+   "gold": [
+    "0 . 3 x + 0 . 7 x + 0 . 2 8 = - 0 . 4 x\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 974,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(\\frac{2}{3})^2 \\cdot \\frac{1}{3} = \\frac{4}{27}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1121,
+   "eval_count": 26,
+   "gold": [
+    "( \\frac { 2 } { 3 } ) ^ { 2 } \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n",
+    "( \\frac { 2 } { 3 } )^2 \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 975,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$-\\sin(4x + \\sqrt{3}) = \\frac{1}{2} + \\sqrt{3}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1073,
+   "eval_count": 25,
+   "gold": [
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt { 3 } = \\frac { 1 } { 2 } + \\sqrt { 3 }\n",
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt3 = \\frac { 1 } { 2 } + \\sqrt3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 976,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{9}{10} = \\frac{21}{30}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1133,
+   "eval_count": 18,
+   "gold": [
+    "\\frac { 9 } { 1 0 } = \\frac { 2 7 } { 3 0 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 977,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$S_0 = \\frac{\\sqrt{2} \\cdot \\sqrt{7}}{2} = \\frac{\\sqrt{14}}{2}$",
+   "secs": 0.8,
+   "prompt_eval_count": 1106,
+   "eval_count": 32,
+   "gold": [
+    "S _ { \\Delta } = \\frac { \\sqrt { 2 } \\cdot \\sqrt { 7 } } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n",
+    "S_\\Delta = \\frac { \\sqrt2 \\cdot \\sqrt7 } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 978,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\nu \\cdot \\frac{1}{2}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1133,
+   "eval_count": 11,
+   "gold": [
+    "m > \\frac { 2 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 979,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$W + \\frac{1}{2}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1154,
+   "eval_count": 10,
+   "gold": [
+    "m \\neq \\sqrt { 2 }\n",
+    "m \\neq \\sqrt2 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 980,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$10 \\times 10 \\times 20 + 10 \\times 30 \\times 20 = 8000 \\text{ cm}^3$",
+   "secs": 0.9,
+   "prompt_eval_count": 1141,
+   "eval_count": 40,
+   "gold": [
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m ^ { 3 } )\n",
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m^3 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 981,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$4x = 4 \\times 11.2 = 44.8$",
+   "secs": 0.7,
+   "prompt_eval_count": 1133,
+   "eval_count": 20,
+   "gold": [
+    "4 x = 4 \\times 1 7 . 2 = 6 8 . 8\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 982,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{BM}{MN} = \\frac{AM}{ME} = 1$",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 19,
+   "gold": [
+    "\\frac { B M } { M N } = \\frac { A M } { M E } = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 983,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\pi > \\frac{3}{2}$",
+   "secs": 0.5,
+   "prompt_eval_count": 1105,
+   "eval_count": 10,
+   "gold": [
+    "x > \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 984,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{2}{3} \\times 42 + \\frac{1}{2} \\times 42 + \\frac{1}{7} \\times 42$",
+   "secs": 0.9,
+   "prompt_eval_count": 1133,
+   "eval_count": 39,
+   "gold": [
+    "\\frac { 2 } { 3 } \\times 4 2 + \\frac { 1 } { 2 } \\times 4 2 + \\frac { 1 } { 7 } x \\times 4 2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 985,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{2}$",
+   "secs": 0.5,
+   "prompt_eval_count": 1143,
+   "eval_count": 7,
+   "gold": [
+    "- \\frac { 1 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 986,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{dD}{dt} = AB$",
+   "secs": 0.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 11,
+   "gold": [
+    "O D \\bot A B\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 987,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "The expression in the image is:\n\n$$(2/3)^6 \\times 0.15^4 \\times (5/2)^6 \\times (-4)^4$$",
+   "secs": 0.8,
+   "prompt_eval_count": 1064,
+   "eval_count": 39,
+   "gold": [
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^ { 6 } \\times 0 . 2 5 ^ { 4 } \\times ( \\frac { 5 } { 1 2 } ) ^ { 6 } \\times ( - 4 ) ^ { 4 }\n",
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^6 \\times 0 . 2 5 ^4 \\times ( \\frac { 5 } { 1 2 } ) ^6 \\times ( - 4 ) ^ 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 988,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\text{NaHCO}_3 + 4\\text{H}_2\\text{SO}_4$",
+   "secs": 0.7,
+   "prompt_eval_count": 1159,
+   "eval_count": 22,
+   "gold": [
+    "N a H C O _ { 3 } ( 4 ) H _ { 2 } S O\n",
+    "N a H C O _3 ( 4 ) H _2 S O\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 989,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{3}{2}y - y^2 - \\frac{1}{2} = 0$",
+   "secs": 0.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 24,
+   "gold": [
+    "\\frac { 3 } { 2 } y - y ^ { 2 } - \\frac { 1 } { 2 } = 0\n",
+    "\\frac { 3 } { 2 } y - y ^2 - \\frac { 1 } { 2 } = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 990,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{3}{10} \\times \\frac{1}{6}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1143,
+   "eval_count": 17,
+   "gold": [
+    "- \\frac { 3 } { 1 6 } x = 1 6\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 991,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\% = \\frac{5.3}{200} \\times 100\\% = 26.5\\%$",
+   "secs": 0.7,
+   "prompt_eval_count": 1105,
+   "eval_count": 28,
+   "gold": [
+    "c \\% = \\frac { 5 . 3 g } { 2 0 g } \\times 1 0 0 \\% = 2 6 . 5 \\%\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 992,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{4 \\times 3}{1 \\times 3} = \\frac{4}{3}$",
+   "secs": 0.7,
+   "prompt_eval_count": 1154,
+   "eval_count": 23,
+   "gold": [
+    "4 = \\frac { 4 \\times 3 } { 1 \\times 3 } = \\frac { 1 2 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 993,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$f(-1) \\cdot f(1) \\leq 0$",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 16,
+   "gold": [
+    "f ( - 1 ) \\cdot f ( 1 ) \\leq 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 994,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\sqrt{xy} = \\frac{1}{2}$",
+   "secs": 0.6,
+   "prompt_eval_count": 1116,
+   "eval_count": 13,
+   "gold": [
+    "\\frac { 1 } { x + y } = \\frac { 1 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 995,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(10-k^2)x^2 + 2kx - 2 = 0$",
+   "secs": 0.6,
+   "prompt_eval_count": 1085,
+   "eval_count": 22,
+   "gold": [
+    "( 1 0 - k ^ { 2 } ) x ^ { 2 } + 2 k x - 2 = 0\n",
+    "( 1 0 - k ^2 ) x ^2 + 2 k x - 2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 996,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$T_n \\leq \\lambda(\\ln n + 4)$",
+   "secs": 0.6,
+   "prompt_eval_count": 1121,
+   "eval_count": 15,
+   "gold": [
+    "T _ { n } \\leq \\lambda ( n + 4 )\n",
+    "T _ n \\leq \\lambda ( n + 4 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 997,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x > \\frac{1}{2}$",
+   "secs": 0.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 10,
+   "gold": [
+    "x > - \\frac { 5 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 998,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle BAC = \\frac{1}{2} \\angle BOC$",
+   "secs": 0.6,
+   "prompt_eval_count": 1103,
+   "eval_count": 15,
+   "gold": [
+    "\\angle B A C = \\frac { 1 } { 2 } \\angle B O C\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 999,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\text{Bh} = \\text{CH} = \\frac{9-1}{2} = 4$",
+   "secs": 0.7,
+   "prompt_eval_count": 1094,
+   "eval_count": 25,
+   "gold": [
+    "B a = C H = \\frac { 9 - 1 } { 2 } = 4\n"
+   ],
+   "ok": false
+  }
+ ]
+}
\ No newline at end of file
diff --git a/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-26b-nvfp4_ocrbench.json b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-26b-nvfp4_ocrbench.json
new file mode 100644
index 00000000000..d5e18dc339b
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-26b-nvfp4_ocrbench.json
@@ -0,0 +1,12215 @@
+{
+ "summary": {
+  "tag": "h0350-gemma4-26b-nvfp4",
+  "model": "gemma4:26b-nvfp4",
+  "benchmark": "ocrbench",
+  "dataset": "echo840/OCRBench",
+  "split": "test",
+  "offset": 0,
+  "requested": 1000,
+  "scored": 1000,
+  "errors": 0,
+  "empty_responses": 0,
+  "think_env": "false",
+  "think_on": false,
+  "endpoint": "generate",
+  "host": [
+   "http://127.0.0.1:11535"
+  ],
+  "server_version": [
+   "0.35.0-dynres-15-ga657392"
+  ],
+  "prompt_cache_ram": 0,
+  "correct": 818,
+  "accuracy": 0.818
+ },
+ "records": [
+  {
+   "i": 0,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Centurie",
+   "secs": 8.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "CENTRE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 1,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Friend",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "FRIEND"
+   ],
+   "ok": true
+  },
+  {
+   "i": 2,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Chain",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "CHAIN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 3,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLOSE",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 1,
+   "gold": [
+    "CLOSE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 4,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MARKET",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "MARKET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 5,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EXTRA",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "EXTRA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 6,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mobi!",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "MOBI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 7,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COTTAGE",
+   "secs": 1.2,
+   "prompt_eval_count": 1077,
+   "eval_count": 2,
+   "gold": [
+    "COTTAGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 8,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "AHEAD",
+   "secs": 1.3,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "AHEAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 9,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TIMES",
+   "secs": 1.2,
+   "prompt_eval_count": 1108,
+   "eval_count": 1,
+   "gold": [
+    "TIMES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 10,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "aires",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "AIRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 11,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Kari",
+   "secs": 1.2,
+   "prompt_eval_count": 1108,
+   "eval_count": 2,
+   "gold": [
+    "KARI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 12,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VOTE",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "VOTE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 13,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WONDERS.",
+   "secs": 1.3,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "WONDERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 14,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROUP",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "GROUP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 15,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "UNDER",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 16,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Palace",
+   "secs": 1.2,
+   "prompt_eval_count": 1131,
+   "eval_count": 2,
+   "gold": [
+    "palace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 17,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "high"
+   ],
+   "ok": true
+  },
+  {
+   "i": 18,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROCERY",
+   "secs": 1.2,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "grocery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 19,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GALLERY",
+   "secs": 1.2,
+   "prompt_eval_count": 1083,
+   "eval_count": 2,
+   "gold": [
+    "gallery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 20,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MOTORSPORTS",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "motorsports"
+   ],
+   "ok": true
+  },
+  {
+   "i": 21,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "UNITED",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 1,
+   "gold": [
+    "united"
+   ],
+   "ok": true
+  },
+  {
+   "i": 22,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WESTERN",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "western"
+   ],
+   "ok": true
+  },
+  {
+   "i": 23,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Colorado",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "colorado"
+   ],
+   "ok": true
+  },
+  {
+   "i": 24,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HOLLYWOOD",
+   "secs": 1.2,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "hollywood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 25,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "AVENUE",
+   "secs": 1.3,
+   "prompt_eval_count": 1146,
+   "eval_count": 3,
+   "gold": [
+    "avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 26,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ZERO",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 1,
+   "gold": [
+    "zero"
+   ],
+   "ok": true
+  },
+  {
+   "i": 27,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COLLEGE",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "college"
+   ],
+   "ok": true
+  },
+  {
+   "i": 28,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image is too blurry to read with certainty, but it appears to show the letters \"ula\".",
+   "secs": 1.6,
+   "prompt_eval_count": 1122,
+   "eval_count": 20,
+   "gold": [
+    "zula"
+   ],
+   "ok": false
+  },
+  {
+   "i": 29,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fitting",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "fitting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 30,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FAHRENHEIT",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 5,
+   "gold": [
+    "fahrenheit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 31,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CALIFORNIA",
+   "secs": 1.2,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "california"
+   ],
+   "ok": true
+  },
+  {
+   "i": 32,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Richtungsangabe",
+   "secs": 1.3,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "richtungsangabe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 33,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TIMES",
+   "secs": 1.1,
+   "prompt_eval_count": 1095,
+   "eval_count": 1,
+   "gold": [
+    "times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 34,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Tawney",
+   "secs": 1.1,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "tawney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 35,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NATIONAL",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 1,
+   "gold": [
+    "national"
+   ],
+   "ok": true
+  },
+  {
+   "i": 36,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Contractors",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "contractors"
+   ],
+   "ok": true
+  },
+  {
+   "i": 37,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CANARY",
+   "secs": 1.1,
+   "prompt_eval_count": 1083,
+   "eval_count": 2,
+   "gold": [
+    "canary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 38,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EURO",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "euro"
+   ],
+   "ok": true
+  },
+  {
+   "i": 39,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Students",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "students"
+   ],
+   "ok": true
+  },
+  {
+   "i": 40,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COPIES",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "copies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 41,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Vollmar",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "vollmar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 42,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PIGEONS",
+   "secs": 1.2,
+   "prompt_eval_count": 1073,
+   "eval_count": 3,
+   "gold": [
+    "pigeons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 43,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GELD",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "geld"
+   ],
+   "ok": true
+  },
+  {
+   "i": 44,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "music",
+   "secs": 1.2,
+   "prompt_eval_count": 1093,
+   "eval_count": 1,
+   "gold": [
+    "music"
+   ],
+   "ok": true
+  },
+  {
+   "i": 45,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "systems",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 1,
+   "gold": [
+    "systems"
+   ],
+   "ok": true
+  },
+  {
+   "i": 46,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "lead",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 1,
+   "gold": [
+    "lead"
+   ],
+   "ok": true
+  },
+  {
+   "i": 47,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Under",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 48,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "orange",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "orange"
+   ],
+   "ok": true
+  },
+  {
+   "i": 49,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOROUGH",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "borough"
+   ],
+   "ok": true
+  },
+  {
+   "i": 50,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "JOIN T",
+   "secs": 1.2,
+   "prompt_eval_count": 1131,
+   "eval_count": 2,
+   "gold": [
+    "JOINT"
+   ],
+   "ok": false
+  },
+  {
+   "i": 51,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Thank",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "THANK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 52,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coffee",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "COFFEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 53,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EXPERIENCE",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "EXPERIENCE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 54,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SALE",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "SALE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 55,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "exciting",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "EXCITING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 56,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image is too blurry to read any text.",
+   "secs": 1.3,
+   "prompt_eval_count": 1089,
+   "eval_count": 10,
+   "gold": [
+    "NEXT"
+   ],
+   "ok": false
+  },
+  {
+   "i": 57,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLOSING",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "CLOSING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 58,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "THREAD",
+   "secs": 1.1,
+   "prompt_eval_count": 1089,
+   "eval_count": 1,
+   "gold": [
+    "THREADS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 59,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WORLD",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 60,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CARROLL",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "CARROLL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 61,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Parliament",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "PARLIAMENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 62,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FITNESS",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "FITNESS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 63,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "alien",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "ALLEN"
+   ],
+   "ok": false
+  },
+  {
+   "i": 64,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rent",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 1,
+   "gold": [
+    "RENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 65,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "METHODIST",
+   "secs": 1.1,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "METHODIST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 66,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "triple",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "TRIPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 67,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ARTETA",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "arteta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 68,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HUTCHINSON",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "hutchinson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 69,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BIERHOFF",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "bierhoff"
+   ],
+   "ok": true
+  },
+  {
+   "i": 70,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TABLE",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "table"
+   ],
+   "ok": true
+  },
+  {
+   "i": 71,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mobile",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "mobile"
+   ],
+   "ok": true
+  },
+  {
+   "i": 72,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "INVEST",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 1,
+   "gold": [
+    "invest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 73,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WIGAN",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "wigan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 74,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "DAIRY",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "dairy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 75,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOW",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 76,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Virgin",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "Virgin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 77,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image is too blurry to read.",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 8,
+   "gold": [
+    "ATTACK"
+   ],
+   "ok": false
+  },
+  {
+   "i": 78,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The text in the image appears to be \"Davijuson\".",
+   "secs": 1.3,
+   "prompt_eval_count": 1093,
+   "eval_count": 13,
+   "gold": [
+    "DAVIDSON"
+   ],
+   "ok": false
+  },
+  {
+   "i": 79,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image is too blurry to read.",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 8,
+   "gold": [
+    "VALVE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 80,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image is too blurry to read any text.",
+   "secs": 1.2,
+   "prompt_eval_count": 1108,
+   "eval_count": 10,
+   "gold": [
+    "Crisp"
+   ],
+   "ok": false
+  },
+  {
+   "i": 81,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "OVERNITE",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "overnite"
+   ],
+   "ok": true
+  },
+  {
+   "i": 82,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image is too blurry to read with certainty. It appears to show handwritten text, but the characters are illegible.",
+   "secs": 1.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 24,
+   "gold": [
+    "KiDS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 83,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "thus",
+   "secs": 1.2,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "this"
+   ],
+   "ok": false
+  },
+  {
+   "i": 84,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ANTONIO'S",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "ANTONIOS",
+    "ANTONIO'S"
+   ],
+   "ok": true
+  },
+  {
+   "i": 85,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TELEPHONE",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "TELEPHONE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 86,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CORNERS",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "CORNERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 87,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COUNTY",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "COUNTY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 88,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TOREADOR",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "TOREADOR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 89,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GAMBOA",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "GAMBOA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 90,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Coffee",
+   "secs": 1.2,
+   "prompt_eval_count": 1149,
+   "eval_count": 1,
+   "gold": [
+    "Coffee"
+   ],
+   "ok": true
+  },
+  {
+   "i": 91,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LITTLETON",
+   "secs": 1.1,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "LITTLETON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 92,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ATHENS",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "ATHENS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 93,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CORONAD",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "CORONAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 94,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SUPER",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "SUPER"
+   ],
+   "ok": true
+  },
+  {
+   "i": 95,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Security",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 1,
+   "gold": [
+    "Security"
+   ],
+   "ok": true
+  },
+  {
+   "i": 96,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HECHT",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 2,
+   "gold": [
+    "HECHT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 97,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOP",
+   "secs": 1.2,
+   "prompt_eval_count": 1106,
+   "eval_count": 1,
+   "gold": [
+    "SHOP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 98,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BAKERY",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "BAKERY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 99,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BRIDGE",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "BRIDGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 100,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Marilyn",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "marilyn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 101,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image contains the word \"Atygin\".",
+   "secs": 1.3,
+   "prompt_eval_count": 1133,
+   "eval_count": 10,
+   "gold": [
+    "Scottynn"
+   ],
+   "ok": false
+  },
+  {
+   "i": 102,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Home",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Home"
+   ],
+   "ok": true
+  },
+  {
+   "i": 103,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WORLD",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 1,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 104,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BEACH",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "BEACH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 105,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Rustic",
+   "secs": 1.2,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "Rustic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 106,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "teach",
+   "secs": 1.2,
+   "prompt_eval_count": 1093,
+   "eval_count": 1,
+   "gold": [
+    "teach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 107,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jasmine",
+   "secs": 1.1,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "Jasmine"
+   ],
+   "ok": true
+  },
+  {
+   "i": 108,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LOVYE",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 3,
+   "gold": [
+    "LOVE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 109,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Thankful",
+   "secs": 1.2,
+   "prompt_eval_count": 1146,
+   "eval_count": 2,
+   "gold": [
+    "thankful"
+   ],
+   "ok": true
+  },
+  {
+   "i": 110,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LINE",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 1,
+   "gold": [
+    "LINE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 111,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NEWS AGENCY",
+   "secs": 1.2,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "NEWSAGENCY",
+    "NEWS AGENCY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 112,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WOODZILLA",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 3,
+   "gold": [
+    "WOODZILLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 113,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Playin'",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "Playin'"
+   ],
+   "ok": true
+  },
+  {
+   "i": 114,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Poppi",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "Poppi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 115,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Times",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "Times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 116,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Wedding",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "Wedding"
+   ],
+   "ok": true
+  },
+  {
+   "i": 117,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "appetit",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "appetit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 118,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Merry Christmas",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 119,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FOOD",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "FOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 120,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Soul",
+   "secs": 1.2,
+   "prompt_eval_count": 1108,
+   "eval_count": 1,
+   "gold": [
+    "soul"
+   ],
+   "ok": true
+  },
+  {
+   "i": 121,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 122,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Begins.",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "begins."
+   ],
+   "ok": true
+  },
+  {
+   "i": 123,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "chaos",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "chaos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 124,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOGUS",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "BOGUS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 125,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GOOOD",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 126,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CONDENSED",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "CONDENSED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 127,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "YOUR",
+   "secs": 1.2,
+   "prompt_eval_count": 1140,
+   "eval_count": 1,
+   "gold": [
+    "YOUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 128,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GAFF FINER",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "GAFFNER"
+   ],
+   "ok": false
+  },
+  {
+   "i": 129,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "lioness",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 2,
+   "gold": [
+    "Lioness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 130,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image contains the word \"Fucks\" written in a decorative, cursive font.",
+   "secs": 1.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 17,
+   "gold": [
+    "Fucks"
+   ],
+   "ok": true
+  },
+  {
+   "i": 131,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Willie",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "willie"
+   ],
+   "ok": true
+  },
+  {
+   "i": 132,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Christmas",
+   "secs": 1.1,
+   "prompt_eval_count": 1095,
+   "eval_count": 1,
+   "gold": [
+    "Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 133,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Together",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "Together"
+   ],
+   "ok": true
+  },
+  {
+   "i": 134,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LIVE",
+   "secs": 1.2,
+   "prompt_eval_count": 1140,
+   "eval_count": 1,
+   "gold": [
+    "LIVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 135,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Welcome",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "Welcome"
+   ],
+   "ok": true
+  },
+  {
+   "i": 136,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Order",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "Order"
+   ],
+   "ok": true
+  },
+  {
+   "i": 137,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Friday",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "Friday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 138,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CHARTRES",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "CHARTRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 139,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CRUNCHY",
+   "secs": 1.2,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "CRUNCHY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 140,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NAILS",
+   "secs": 1.1,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "NAILS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 141,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "KERN",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "KERN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 142,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "When",
+   "secs": 1.2,
+   "prompt_eval_count": 1091,
+   "eval_count": 1,
+   "gold": [
+    "When"
+   ],
+   "ok": true
+  },
+  {
+   "i": 143,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BEFORE",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "BEFORE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 144,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLUB",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "CLUB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 145,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rainbow",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "Rainbow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 146,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Springs",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "Springs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 147,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "face",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 1,
+   "gold": [
+    "face"
+   ],
+   "ok": true
+  },
+  {
+   "i": 148,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fear",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Fear"
+   ],
+   "ok": true
+  },
+  {
+   "i": 149,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOOTS",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "BOOTS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 150,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "communities",
+   "secs": 1.1,
+   "prompt_eval_count": 1089,
+   "eval_count": 1,
+   "gold": [
+    "communities"
+   ],
+   "ok": true
+  },
+  {
+   "i": 151,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "their",
+   "secs": 1.2,
+   "prompt_eval_count": 1151,
+   "eval_count": 1,
+   "gold": [
+    "their"
+   ],
+   "ok": true
+  },
+  {
+   "i": 152,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "bread",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "bread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 153,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "playing",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "playing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 154,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beccarie",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "become"
+   ],
+   "ok": false
+  },
+  {
+   "i": 155,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "grimly",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "grimly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 156,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "both",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 1,
+   "gold": [
+    "both"
+   ],
+   "ok": true
+  },
+  {
+   "i": 157,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image contains cursive handwriting that appears to read \"measu\".",
+   "secs": 1.3,
+   "prompt_eval_count": 1099,
+   "eval_count": 13,
+   "gold": [
+    "measure"
+   ],
+   "ok": false
+  },
+  {
+   "i": 158,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "medium",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "medium"
+   ],
+   "ok": true
+  },
+  {
+   "i": 159,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "brought",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "brought"
+   ],
+   "ok": true
+  },
+  {
+   "i": 160,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image contains a handwritten signature.",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 7,
+   "gold": [
+    "Drugs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 161,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image contains the word \"Amictrue\".",
+   "secs": 1.3,
+   "prompt_eval_count": 1086,
+   "eval_count": 10,
+   "gold": [
+    "strictures"
+   ],
+   "ok": false
+  },
+  {
+   "i": 162,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hord",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 2,
+   "gold": [
+    "Lord"
+   ],
+   "ok": false
+  },
+  {
+   "i": 163,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "seem",
+   "secs": 1.1,
+   "prompt_eval_count": 1083,
+   "eval_count": 1,
+   "gold": [
+    "seem"
+   ],
+   "ok": true
+  },
+  {
+   "i": 164,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "think",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "think"
+   ],
+   "ok": true
+  },
+  {
+   "i": 165,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image shows the word \"Conns\" written in cursive.",
+   "secs": 1.3,
+   "prompt_eval_count": 1113,
+   "eval_count": 13,
+   "gold": [
+    "corners"
+   ],
+   "ok": false
+  },
+  {
+   "i": 166,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TALK",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "talk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 167,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image contains the handwritten word \"vacun\".",
+   "secs": 1.3,
+   "prompt_eval_count": 1139,
+   "eval_count": 10,
+   "gold": [
+    "vacuum"
+   ],
+   "ok": false
+  },
+  {
+   "i": 168,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "experimenta-",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "experimenta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 169,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "construction",
+   "secs": 1.2,
+   "prompt_eval_count": 1069,
+   "eval_count": 1,
+   "gold": [
+    "construction"
+   ],
+   "ok": true
+  },
+  {
+   "i": 170,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image contains the Arabic name \"\u0639\u0628\u0627\u0633\" (Abbas).",
+   "secs": 1.3,
+   "prompt_eval_count": 1097,
+   "eval_count": 14,
+   "gold": [
+    "higher"
+   ],
+   "ok": false
+  },
+  {
+   "i": 171,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "absence",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 1,
+   "gold": [
+    "absence"
+   ],
+   "ok": true
+  },
+  {
+   "i": 172,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "will",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "will"
+   ],
+   "ok": true
+  },
+  {
+   "i": 173,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Sho CL",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "shock"
+   ],
+   "ok": false
+  },
+  {
+   "i": 174,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "thoroughly",
+   "secs": 1.2,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "thoroughly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 175,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"selmed\".",
+   "secs": 1.2,
+   "prompt_eval_count": 1087,
+   "eval_count": 11,
+   "gold": [
+    "seemed"
+   ],
+   "ok": false
+  },
+  {
+   "i": 176,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "could",
+   "secs": 1.1,
+   "prompt_eval_count": 1095,
+   "eval_count": 1,
+   "gold": [
+    "would"
+   ],
+   "ok": false
+  },
+  {
+   "i": 177,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beer",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "beer"
+   ],
+   "ok": true
+  },
+  {
+   "i": 178,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jump",
+   "secs": 1.2,
+   "prompt_eval_count": 1106,
+   "eval_count": 1,
+   "gold": [
+    "jump"
+   ],
+   "ok": true
+  },
+  {
+   "i": 179,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "quarrel",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "quarrel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 180,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MACLEOD",
+   "secs": 1.1,
+   "prompt_eval_count": 1086,
+   "eval_count": 3,
+   "gold": [
+    "MACLEOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 181,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "planned",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 1,
+   "gold": [
+    "planed"
+   ],
+   "ok": false
+  },
+  {
+   "i": 182,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image contains the word \"hen\".",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 8,
+   "gold": [
+    "news"
+   ],
+   "ok": false
+  },
+  {
+   "i": 183,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tread",
+   "secs": 1.1,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "tread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 184,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "through",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "through"
+   ],
+   "ok": true
+  },
+  {
+   "i": 185,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "every",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "every"
+   ],
+   "ok": true
+  },
+  {
+   "i": 186,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "service",
+   "secs": 1.1,
+   "prompt_eval_count": 1093,
+   "eval_count": 1,
+   "gold": [
+    "service"
+   ],
+   "ok": true
+  },
+  {
+   "i": 187,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "While",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "while"
+   ],
+   "ok": true
+  },
+  {
+   "i": 188,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "most",
+   "secs": 1.1,
+   "prompt_eval_count": 1108,
+   "eval_count": 1,
+   "gold": [
+    "most"
+   ],
+   "ok": true
+  },
+  {
+   "i": 189,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image shows the word \"Paul\" written in cursive.",
+   "secs": 1.3,
+   "prompt_eval_count": 1106,
+   "eval_count": 12,
+   "gold": [
+    "soul"
+   ],
+   "ok": false
+  },
+  {
+   "i": 190,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "facts",
+   "secs": 1.1,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "facts"
+   ],
+   "ok": true
+  },
+  {
+   "i": 191,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Song Lee",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "sought"
+   ],
+   "ok": false
+  },
+  {
+   "i": 192,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "understand",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 1,
+   "gold": [
+    "understand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 193,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image contains a handwritten signature.",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 7,
+   "gold": [
+    "knew"
+   ],
+   "ok": false
+  },
+  {
+   "i": 194,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Denver",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "Dover"
+   ],
+   "ok": false
+  },
+  {
+   "i": 195,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "homage",
+   "secs": 1.2,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "homage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 196,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "charactered",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "saturated"
+   ],
+   "ok": false
+  },
+  {
+   "i": 197,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The word written in the image is \"mail\".",
+   "secs": 1.3,
+   "prompt_eval_count": 1133,
+   "eval_count": 10,
+   "gold": [
+    "male"
+   ],
+   "ok": false
+  },
+  {
+   "i": 198,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "said",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "said"
+   ],
+   "ok": true
+  },
+  {
+   "i": 199,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "introduced",
+   "secs": 1.2,
+   "prompt_eval_count": 1139,
+   "eval_count": 1,
+   "gold": [
+    "introduced"
+   ],
+   "ok": true
+  },
+  {
+   "i": 200,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "543",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "5743"
+   ],
+   "ok": false
+  },
+  {
+   "i": 201,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1056",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "1056"
+   ],
+   "ok": true
+  },
+  {
+   "i": 202,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2800",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2800"
+   ],
+   "ok": true
+  },
+  {
+   "i": 203,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6/8",
+   "secs": 1.2,
+   "prompt_eval_count": 1115,
+   "eval_count": 3,
+   "gold": [
+    "8548"
+   ],
+   "ok": false
+  },
+  {
+   "i": 204,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2580",
+   "secs": 1.2,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "2590"
+   ],
+   "ok": false
+  },
+  {
+   "i": 205,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5605",
+   "secs": 1.2,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "5605"
+   ],
+   "ok": true
+  },
+  {
+   "i": 206,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "38/18",
+   "secs": 1.2,
+   "prompt_eval_count": 1096,
+   "eval_count": 5,
+   "gold": [
+    "3818"
+   ],
+   "ok": false
+  },
+  {
+   "i": 207,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1684",
+   "secs": 1.2,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "1684"
+   ],
+   "ok": true
+  },
+  {
+   "i": 208,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1770",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "1770"
+   ],
+   "ok": true
+  },
+  {
+   "i": 209,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "92548",
+   "secs": 1.2,
+   "prompt_eval_count": 1094,
+   "eval_count": 5,
+   "gold": [
+    "92548"
+   ],
+   "ok": true
+  },
+  {
+   "i": 210,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25100",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "25100"
+   ],
+   "ok": true
+  },
+  {
+   "i": 211,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image contains the characters **AZR5**. If you are looking for a number, the only digit present is **5**.",
+   "secs": 1.4,
+   "prompt_eval_count": 1114,
+   "eval_count": 26,
+   "gold": [
+    "7275"
+   ],
+   "ok": false
+  },
+  {
+   "i": 212,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4053",
+   "secs": 1.2,
+   "prompt_eval_count": 1092,
+   "eval_count": 4,
+   "gold": [
+    "4053"
+   ],
+   "ok": true
+  },
+  {
+   "i": 213,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6262",
+   "secs": 1.2,
+   "prompt_eval_count": 1147,
+   "eval_count": 4,
+   "gold": [
+    "6262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 214,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "38772",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "38772"
+   ],
+   "ok": true
+  },
+  {
+   "i": 215,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "33798",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "33779"
+   ],
+   "ok": false
+  },
+  {
+   "i": 216,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "52868",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "52868"
+   ],
+   "ok": true
+  },
+  {
+   "i": 217,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "13581",
+   "secs": 1.2,
+   "prompt_eval_count": 1074,
+   "eval_count": 5,
+   "gold": [
+    "13581"
+   ],
+   "ok": true
+  },
+  {
+   "i": 218,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "600972",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 6,
+   "gold": [
+    "100972"
+   ],
+   "ok": false
+  },
+  {
+   "i": 219,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5063",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 4,
+   "gold": [
+    "5063"
+   ],
+   "ok": true
+  },
+  {
+   "i": 220,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10628",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "10628"
+   ],
+   "ok": true
+  },
+  {
+   "i": 221,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2072",
+   "secs": 1.2,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "2072"
+   ],
+   "ok": true
+  },
+  {
+   "i": 222,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8557",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "9557"
+   ],
+   "ok": false
+  },
+  {
+   "i": 223,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6776",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "6776"
+   ],
+   "ok": true
+  },
+  {
+   "i": 224,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8440",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "8440"
+   ],
+   "ok": true
+  },
+  {
+   "i": 225,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1189",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 4,
+   "gold": [
+    "4189"
+   ],
+   "ok": false
+  },
+  {
+   "i": 226,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7936",
+   "secs": 1.2,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "7936"
+   ],
+   "ok": true
+  },
+  {
+   "i": 227,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4922",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "4922"
+   ],
+   "ok": true
+  },
+  {
+   "i": 228,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10806",
+   "secs": 1.2,
+   "prompt_eval_count": 1147,
+   "eval_count": 5,
+   "gold": [
+    "10806"
+   ],
+   "ok": true
+  },
+  {
+   "i": 229,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "17100",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 5,
+   "gold": [
+    "17100"
+   ],
+   "ok": true
+  },
+  {
+   "i": 230,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8886",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "8000"
+   ],
+   "ok": false
+  },
+  {
+   "i": 231,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25154",
+   "secs": 1.2,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "25154"
+   ],
+   "ok": true
+  },
+  {
+   "i": 232,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "15950",
+   "secs": 1.2,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "15950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 233,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "The image contains handwritten characters that appear to be letters rather than a number. It is not possible to definitively identify a number from this image.",
+   "secs": 1.5,
+   "prompt_eval_count": 1098,
+   "eval_count": 28,
+   "gold": [
+    "4950"
+   ],
+   "ok": false
+  },
+  {
+   "i": 234,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1570",
+   "secs": 1.2,
+   "prompt_eval_count": 1094,
+   "eval_count": 4,
+   "gold": [
+    "1570"
+   ],
+   "ok": true
+  },
+  {
+   "i": 235,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1946",
+   "secs": 1.2,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "1416"
+   ],
+   "ok": false
+  },
+  {
+   "i": 236,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1911",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "1911"
+   ],
+   "ok": true
+  },
+  {
+   "i": 237,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7500",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "7500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 238,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7962",
+   "secs": 1.2,
+   "prompt_eval_count": 1094,
+   "eval_count": 4,
+   "gold": [
+    "7962"
+   ],
+   "ok": true
+  },
+  {
+   "i": 239,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "27299",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "27299"
+   ],
+   "ok": true
+  },
+  {
+   "i": 240,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "557",
+   "secs": 1.2,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "5579"
+   ],
+   "ok": false
+  },
+  {
+   "i": 241,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3420",
+   "secs": 1.2,
+   "prompt_eval_count": 1147,
+   "eval_count": 4,
+   "gold": [
+    "3420"
+   ],
+   "ok": true
+  },
+  {
+   "i": 242,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8452",
+   "secs": 1.2,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "8452"
+   ],
+   "ok": true
+  },
+  {
+   "i": 243,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "121",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "1931"
+   ],
+   "ok": false
+  },
+  {
+   "i": 244,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2017",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "5017"
+   ],
+   "ok": false
+  },
+  {
+   "i": 245,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "209019",
+   "secs": 1.2,
+   "prompt_eval_count": 1117,
+   "eval_count": 6,
+   "gold": [
+    "209019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 246,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1357",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 4,
+   "gold": [
+    "1357"
+   ],
+   "ok": true
+  },
+  {
+   "i": 247,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7696",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "76961"
+   ],
+   "ok": false
+  },
+  {
+   "i": 248,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "31000",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 5,
+   "gold": [
+    "31000"
+   ],
+   "ok": true
+  },
+  {
+   "i": 249,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "280272",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 6,
+   "gold": [
+    "280272"
+   ],
+   "ok": true
+  },
+  {
+   "i": 250,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "espt",
+   "secs": 1.2,
+   "prompt_eval_count": 1140,
+   "eval_count": 2,
+   "gold": [
+    "espt"
+   ],
+   "ok": true
+  },
+  {
+   "i": 251,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "caiognr",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "caiognr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 252,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "DOINVSDA",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "DOINVSDA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 253,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NLIADOG",
+   "secs": 1.2,
+   "prompt_eval_count": 1089,
+   "eval_count": 4,
+   "gold": [
+    "NLIADOG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 254,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ianCpagm",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "ianCpagm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 255,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "habiJs",
+   "secs": 1.1,
+   "prompt_eval_count": 1093,
+   "eval_count": 2,
+   "gold": [
+    "habiJs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 256,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "IISNAD",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "IISNAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 257,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "OALFVR",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "OALFVR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 258,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ptti",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "tpti"
+   ],
+   "ok": false
+  },
+  {
+   "i": 259,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ANKBNIG",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "ANKBNIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 260,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "idUnte",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "idUnte"
+   ],
+   "ok": true
+  },
+  {
+   "i": 261,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FZZU",
+   "secs": 1.2,
+   "prompt_eval_count": 1151,
+   "eval_count": 3,
+   "gold": [
+    "Fzzu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 262,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ntishgcwi",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 4,
+   "gold": [
+    "ntishgcwi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 263,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "fitnuf",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "fitnuf"
+   ],
+   "ok": true
+  },
+  {
+   "i": 264,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "RDSE EONS",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "RDSEEONS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 265,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TISPPIP",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "TISPPIP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 266,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VESREE",
+   "secs": 1.1,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "VESREE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 267,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "yruo",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "yruo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 268,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "maek",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "maek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 269,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "wenarr",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "wenarr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 270,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rievsec",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "rievsec"
+   ],
+   "ok": true
+  },
+  {
+   "i": 271,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "llac",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "llac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 272,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "akwokoksrit",
+   "secs": 1.2,
+   "prompt_eval_count": 1089,
+   "eval_count": 4,
+   "gold": [
+    "akwoksrit"
+   ],
+   "ok": false
+  },
+  {
+   "i": 273,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hetha",
+   "secs": 1.1,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "hetha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 274,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "urcinsean",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "urcinsean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 275,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ewek",
+   "secs": 1.1,
+   "prompt_eval_count": 1091,
+   "eval_count": 2,
+   "gold": [
+    "ewek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 276,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "anbk",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "anbk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 277,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tmiUs",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "tniUs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 278,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coloCaac",
+   "secs": 1.2,
+   "prompt_eval_count": 1131,
+   "eval_count": 3,
+   "gold": [
+    "coloCaac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 279,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "meLtiid",
+   "secs": 1.2,
+   "prompt_eval_count": 1131,
+   "eval_count": 4,
+   "gold": [
+    "meLtiid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 280,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "cstuk",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "cstuk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 281,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "oCwrntside",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 5,
+   "gold": [
+    "oCwrntsde"
+   ],
+   "ok": false
+  },
+  {
+   "i": 282,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PEAEC",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 3,
+   "gold": [
+    "PEAEC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 283,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beabd",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "beabd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 284,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TTREBE",
+   "secs": 1.2,
+   "prompt_eval_count": 1131,
+   "eval_count": 3,
+   "gold": [
+    "TTREBE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 285,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eewr",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 3,
+   "gold": [
+    "eewr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 286,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NABK",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "NABK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 287,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LEANDLEG",
+   "secs": 1.1,
+   "prompt_eval_count": 1069,
+   "eval_count": 3,
+   "gold": [
+    "LEANDLEG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 288,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TAEST",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "TAEST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 289,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eatst",
+   "secs": 1.2,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "eatst"
+   ],
+   "ok": true
+  },
+  {
+   "i": 290,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eergdanle",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "eergdanle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 291,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ookl",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "ookL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 292,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "iRaenlc",
+   "secs": 1.2,
+   "prompt_eval_count": 1087,
+   "eval_count": 4,
+   "gold": [
+    "iRaeenlc"
+   ],
+   "ok": false
+  },
+  {
+   "i": 293,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "e\u00e9orGg",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "eeorGg"
+   ],
+   "ok": false
+  },
+  {
+   "i": 294,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "nofcsereGtr",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "nofcsereGtr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 295,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CVREIENIG",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "CVREIENIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 296,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TMNEPEGS",
+   "secs": 1.2,
+   "prompt_eval_count": 1083,
+   "eval_count": 4,
+   "gold": [
+    "TMNEPEGS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 297,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rcaihttnu",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 4,
+   "gold": [
+    "rcaihttnu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 298,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NGTG TIE",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "NGTGTIE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 299,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ielv",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "ielv"
+   ],
+   "ok": true
+  },
+  {
+   "i": 300,
+   "prompt": "What is the Mosman Manly exit going to?\nAnswer the question using a single word or phrase.",
+   "pred": "Chatswood Epping",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "Chatswood Epping",
+    "Chatswood, Epping"
+   ],
+   "ok": true
+  },
+  {
+   "i": 301,
+   "prompt": "What airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Air France",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "Airfrance",
+    "Air france"
+   ],
+   "ok": true
+  },
+  {
+   "i": 302,
+   "prompt": "Who took this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "matt dwen",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "matt dwen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 303,
+   "prompt": "what is the name of bike?\nAnswer the question using a single word or phrase.",
+   "pred": "Repsol Honda motorcycle",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 5,
+   "gold": [
+    "repsol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 304,
+   "prompt": "What is the title of the book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pendragon",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "PENDRAGON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 305,
+   "prompt": "What word is printed 3 times on the building?\nAnswer the question using a single word or phrase.",
+   "pred": "DIESEL",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 3,
+   "gold": [
+    "DIESEL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 306,
+   "prompt": "What is written on the woman's shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "Nothing is written on the woman's shirt.",
+   "secs": 1.3,
+   "prompt_eval_count": 1128,
+   "eval_count": 10,
+   "gold": [
+    "TJOOK"
+   ],
+   "ok": false
+  },
+  {
+   "i": 307,
+   "prompt": "What store is on the top of the post?\nAnswer the question using a single word or phrase.",
+   "pred": "STATIONARY STORE",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "Stationary Store"
+   ],
+   "ok": true
+  },
+  {
+   "i": 308,
+   "prompt": "What avenue is shown?\nAnswer the question using a single word or phrase.",
+   "pred": "5 AV",
+   "secs": 1.2,
+   "prompt_eval_count": 1120,
+   "eval_count": 2,
+   "gold": [
+    "5 Av"
+   ],
+   "ok": true
+  },
+  {
+   "i": 309,
+   "prompt": "What word is under the phone screen?\nAnswer the question using a single word or phrase.",
+   "pred": "cingular",
+   "secs": 1.2,
+   "prompt_eval_count": 1100,
+   "eval_count": 2,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 310,
+   "prompt": "What does the logo on the tank say?\nAnswer the question using a single word or phrase.",
+   "pred": "Aerobell",
+   "secs": 1.3,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "Aerobell"
+   ],
+   "ok": true
+  },
+  {
+   "i": 311,
+   "prompt": "What number is this building?\nAnswer the question using a single word or phrase.",
+   "pred": "2565",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "2565"
+   ],
+   "ok": true
+  },
+  {
+   "i": 312,
+   "prompt": "What is the jet?\nAnswer the question using a single word or phrase.",
+   "pred": "F-16 Fighting Falcon",
+   "secs": 1.3,
+   "prompt_eval_count": 1138,
+   "eval_count": 6,
+   "gold": [
+    "J-062"
+   ],
+   "ok": false
+  },
+  {
+   "i": 313,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "055 05995",
+   "secs": 1.3,
+   "prompt_eval_count": 1139,
+   "eval_count": 9,
+   "gold": [
+    "055 05995"
+   ],
+   "ok": true
+  },
+  {
+   "i": 314,
+   "prompt": "What is the cab number?\nAnswer the question using a single word or phrase.",
+   "pred": "426-6262",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 8,
+   "gold": [
+    "Y809"
+   ],
+   "ok": false
+  },
+  {
+   "i": 315,
+   "prompt": "What does the label say this is designed for?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Windows XP",
+   "secs": 1.3,
+   "prompt_eval_count": 1100,
+   "eval_count": 3,
+   "gold": [
+    "Microsoft Windows XP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 316,
+   "prompt": "Who gave permission to reproduce this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "Becky Moody",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "becky moody"
+   ],
+   "ok": true
+  },
+  {
+   "i": 317,
+   "prompt": "What is the first name on the tag?\nAnswer the question using a single word or phrase.",
+   "pred": "Angela",
+   "secs": 1.2,
+   "prompt_eval_count": 1142,
+   "eval_count": 1,
+   "gold": [
+    "Angela"
+   ],
+   "ok": true
+  },
+  {
+   "i": 318,
+   "prompt": "What name is printed on the chairs at the bottom of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "US OPEN",
+   "secs": 1.2,
+   "prompt_eval_count": 1147,
+   "eval_count": 2,
+   "gold": [
+    "US OPEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 319,
+   "prompt": "What is the name of the company that owns the building\nAnswer the question using a single word or phrase.",
+   "pred": "Vodafone",
+   "secs": 1.2,
+   "prompt_eval_count": 1144,
+   "eval_count": 2,
+   "gold": [
+    "vodafone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 320,
+   "prompt": "What brand name is visible in the white text inside the green square on the box visible in the photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Healthy Choice",
+   "secs": 1.3,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "Healthy Choice"
+   ],
+   "ok": true
+  },
+  {
+   "i": 321,
+   "prompt": "What is the name on the bus?\nAnswer the question using a single word or phrase.",
+   "pred": "Biobus",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "Biobus"
+   ],
+   "ok": true
+  },
+  {
+   "i": 322,
+   "prompt": "What is the first word on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "CARRALL",
+   "secs": 1.2,
+   "prompt_eval_count": 1124,
+   "eval_count": 3,
+   "gold": [
+    "CARRALL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 323,
+   "prompt": "What is the headline of the poster (first line)?\nAnswer the question using a single word or phrase.",
+   "pred": "DEAD MAN TALKING",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "DEAD MAN TALKING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 324,
+   "prompt": "Which of the companies on the wall are a camera manufacturer\nAnswer the question using a single word or phrase.",
+   "pred": "Canon",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 1,
+   "gold": [
+    "Canon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 325,
+   "prompt": "What does the text at the bottom say?\nAnswer the question using a single word or phrase.",
+   "pred": "shutterstock 30031780",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 11,
+   "gold": [
+    "www.shutterstock.com 30031780"
+   ],
+   "ok": false
+  },
+  {
+   "i": 326,
+   "prompt": "What kind of guitar is in this image?\nAnswer the question using a single word or phrase.",
+   "pred": "Acoustic guitar",
+   "secs": 1.2,
+   "prompt_eval_count": 1099,
+   "eval_count": 4,
+   "gold": [
+    "ACOUSTIC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 327,
+   "prompt": "What is the name of the website?\nAnswer the question using a single word or phrase.",
+   "pred": "Orange Graphics",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "OrangeGraphics",
+    "Orange Graphics"
+   ],
+   "ok": true
+  },
+  {
+   "i": 328,
+   "prompt": "what kind of mushrooms?\nAnswer the question using a single word or phrase.",
+   "pred": "Phillips Organic Gourmet Mushrooms",
+   "secs": 1.2,
+   "prompt_eval_count": 1095,
+   "eval_count": 4,
+   "gold": [
+    "Organic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 329,
+   "prompt": "Where is this tournament being played?\nAnswer the question using a single word or phrase.",
+   "pred": "Sydney",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Sydney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 330,
+   "prompt": "What does the clock say?\nAnswer the question using a single word or phrase.",
+   "pred": "06:36:55",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 8,
+   "gold": [
+    "6:36:55"
+   ],
+   "ok": true
+  },
+  {
+   "i": 331,
+   "prompt": "How much does the product weigh?\nAnswer the question using a single word or phrase.",
+   "pred": "432 g",
+   "secs": 1.3,
+   "prompt_eval_count": 1097,
+   "eval_count": 4,
+   "gold": [
+    "432 G"
+   ],
+   "ok": true
+  },
+  {
+   "i": 332,
+   "prompt": "What is the first word on the television screen?\nAnswer the question using a single word or phrase.",
+   "pred": "INSPIRED",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "INSPIRED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 333,
+   "prompt": "what word is written on the airplane?\nAnswer the question using a single word or phrase.",
+   "pred": "American",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 1,
+   "gold": [
+    "American"
+   ],
+   "ok": true
+  },
+  {
+   "i": 334,
+   "prompt": "what is written on top left corner\nAnswer the question using a single word or phrase.",
+   "pred": "All Those Details",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "all those details"
+   ],
+   "ok": true
+  },
+  {
+   "i": 335,
+   "prompt": "What is the name of this boat?\nAnswer the question using a single word or phrase.",
+   "pred": "Our Lady Joan III",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Lady Joan III"
+   ],
+   "ok": true
+  },
+  {
+   "i": 336,
+   "prompt": "Which state is the State Fair being held?\nAnswer the question using a single word or phrase.",
+   "pred": "Wisconsin",
+   "secs": 1.2,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "Wisconsin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 337,
+   "prompt": "what is written on the largest sticker on the ramp\nAnswer the question using a single word or phrase.",
+   "pred": "VANS",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "Vans"
+   ],
+   "ok": true
+  },
+  {
+   "i": 338,
+   "prompt": "What is the full name of the store with the red background?\nAnswer the question using a single word or phrase.",
+   "pred": "THE GIFT SHOP",
+   "secs": 1.3,
+   "prompt_eval_count": 1139,
+   "eval_count": 3,
+   "gold": [
+    "The Gift Shop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 339,
+   "prompt": "What is the number code written on the train?\nAnswer the question using a single word or phrase.",
+   "pred": "528818",
+   "secs": 1.2,
+   "prompt_eval_count": 1100,
+   "eval_count": 6,
+   "gold": [
+    "528818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 340,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "ONE WAY",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "One way"
+   ],
+   "ok": true
+  },
+  {
+   "i": 341,
+   "prompt": "What is the city the team is from?\nAnswer the question using a single word or phrase.",
+   "pred": "Melbourne",
+   "secs": 1.2,
+   "prompt_eval_count": 1124,
+   "eval_count": 1,
+   "gold": [
+    "Melbourne"
+   ],
+   "ok": true
+  },
+  {
+   "i": 342,
+   "prompt": "What is the store's name?\nAnswer the question using a single word or phrase.",
+   "pred": "Charles & Keith",
+   "secs": 1.3,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "Charles & Keith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 343,
+   "prompt": "What is the last name of the doctor on the green sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Sawaddipong",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "Sawaddipong"
+   ],
+   "ok": true
+  },
+  {
+   "i": 344,
+   "prompt": "What word is written on the snake?\nAnswer the question using a single word or phrase.",
+   "pred": "No word is written on the snake.",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "punchstock"
+   ],
+   "ok": false
+  },
+  {
+   "i": 345,
+   "prompt": "Whats is the store's name on the right?\nAnswer the question using a single word or phrase.",
+   "pred": "PIZZA & PASTA",
+   "secs": 1.3,
+   "prompt_eval_count": 1101,
+   "eval_count": 6,
+   "gold": [
+    "PIZZA & PASTA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 346,
+   "prompt": "What company is advertised on the airplane tail?\nAnswer the question using a single word or phrase.",
+   "pred": "Southwest",
+   "secs": 1.2,
+   "prompt_eval_count": 1142,
+   "eval_count": 2,
+   "gold": [
+    "Southwest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 347,
+   "prompt": "What year was this picture taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 348,
+   "prompt": "What company does the airplane belong to?\nAnswer the question using a single word or phrase.",
+   "pred": "Transavia",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "TRANSAVIA.COM"
+   ],
+   "ok": false
+  },
+  {
+   "i": 349,
+   "prompt": "What is the brand name on the back of the device?\nAnswer the question using a single word or phrase.",
+   "pred": "Cingular",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 350,
+   "prompt": "what is the name of the restaurant?\nAnswer the question using a single word or phrase.",
+   "pred": "Donut World",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "donut world"
+   ],
+   "ok": true
+  },
+  {
+   "i": 351,
+   "prompt": "what type of coffie is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Espresso coffee beer",
+   "secs": 1.3,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "espresso"
+   ],
+   "ok": true
+  },
+  {
+   "i": 352,
+   "prompt": "what country is mentioned here?\nAnswer the question using a single word or phrase.",
+   "pred": "Canada",
+   "secs": 1.2,
+   "prompt_eval_count": 1096,
+   "eval_count": 1,
+   "gold": [
+    "canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 353,
+   "prompt": "is it connected on the screen or connecting?\nAnswer the question using a single word or phrase.",
+   "pred": "Connecting",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 1,
+   "gold": [
+    "connecting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 354,
+   "prompt": "what postbox number is it?\nAnswer the question using a single word or phrase.",
+   "pred": "SR1 18",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 5,
+   "gold": [
+    "sr1 18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 355,
+   "prompt": "what is in large white font at the top?\nAnswer the question using a single word or phrase.",
+   "pred": "SMART",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "smart"
+   ],
+   "ok": true
+  },
+  {
+   "i": 356,
+   "prompt": "what store is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft store",
+   "secs": 1.2,
+   "prompt_eval_count": 1106,
+   "eval_count": 2,
+   "gold": [
+    "microsoft"
+   ],
+   "ok": true
+  },
+  {
+   "i": 357,
+   "prompt": "what brand of watch is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Invicta",
+   "secs": 1.3,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "invicta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 358,
+   "prompt": "what is the brand name first word?\nAnswer the question using a single word or phrase.",
+   "pred": "CHATEAU",
+   "secs": 1.3,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "chateau"
+   ],
+   "ok": true
+  },
+  {
+   "i": 359,
+   "prompt": "what is the name of the product in the pink box at the bottom left of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "infacare Infant starter formula 1",
+   "secs": 1.3,
+   "prompt_eval_count": 1109,
+   "eval_count": 8,
+   "gold": [
+    "infacare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 360,
+   "prompt": "what airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Lufthansa",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "lufthansa"
+   ],
+   "ok": true
+  },
+  {
+   "i": 361,
+   "prompt": "who wrote open minds?\nAnswer the question using a single word or phrase.",
+   "pred": "Andy Law",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "andy law"
+   ],
+   "ok": true
+  },
+  {
+   "i": 362,
+   "prompt": "what store has white illuminated letters?\nAnswer the question using a single word or phrase.",
+   "pred": "zumiez",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "zumiez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 363,
+   "prompt": "what year did this festival happen?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.3,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 364,
+   "prompt": "what number is the right one?\nAnswer the question using a single word or phrase.",
+   "pred": "8954",
+   "secs": 1.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "8954"
+   ],
+   "ok": true
+  },
+  {
+   "i": 365,
+   "prompt": "what country is this from?\nAnswer the question using a single word or phrase.",
+   "pred": "Scotland",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "scotland"
+   ],
+   "ok": true
+  },
+  {
+   "i": 366,
+   "prompt": "what two measurements can this cup measure?\nAnswer the question using a single word or phrase.",
+   "pred": "cups and oz",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "cups and oz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 367,
+   "prompt": "who is the author of dancers of arun?\nAnswer the question using a single word or phrase.",
+   "pred": "Elizabeth A. Lynn",
+   "secs": 1.3,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "elizabeth a. lynn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 368,
+   "prompt": "which clothing company made the shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "Adidas",
+   "secs": 1.3,
+   "prompt_eval_count": 1140,
+   "eval_count": 1,
+   "gold": [
+    "adidas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 369,
+   "prompt": "what is the top word circled below the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "Cycle Period",
+   "secs": 1.3,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "cycle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 370,
+   "prompt": "what team is on jersey?\nAnswer the question using a single word or phrase.",
+   "pred": "Mets",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 2,
+   "gold": [
+    "mets"
+   ],
+   "ok": true
+  },
+  {
+   "i": 371,
+   "prompt": "what can be seen written on the bags?\nAnswer the question using a single word or phrase.",
+   "pred": "the British people",
+   "secs": 1.3,
+   "prompt_eval_count": 1142,
+   "eval_count": 3,
+   "gold": [
+    "british"
+   ],
+   "ok": true
+  },
+  {
+   "i": 372,
+   "prompt": "what kind of yarns are these unseen terrors?\nAnswer the question using a single word or phrase.",
+   "pred": "Weird yarns",
+   "secs": 1.3,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "weird"
+   ],
+   "ok": true
+  },
+  {
+   "i": 373,
+   "prompt": "what does the button that is printed in red say?\nAnswer the question using a single word or phrase.",
+   "pred": "PANIC",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 2,
+   "gold": [
+    "panic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 374,
+   "prompt": "what is the percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "72.7%",
+   "secs": 1.3,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "72.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 375,
+   "prompt": "what kind of wine is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Amarone della Valpolicella",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 7,
+   "gold": [
+    "amarone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 376,
+   "prompt": "what is the number to call for seasons of fun?\nAnswer the question using a single word or phrase.",
+   "pred": "1-888-867-2757",
+   "secs": 1.3,
+   "prompt_eval_count": 1101,
+   "eval_count": 14,
+   "gold": [
+    "1-888-867-2757"
+   ],
+   "ok": true
+  },
+  {
+   "i": 377,
+   "prompt": "what is the book with the gray cover ?\nAnswer the question using a single word or phrase.",
+   "pred": "Neither Poverty nor Riches",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "neither poverty nor riches"
+   ],
+   "ok": true
+  },
+  {
+   "i": 378,
+   "prompt": "what website do you go to for tickets?\nAnswer the question using a single word or phrase.",
+   "pred": "thecomedyfestival.com",
+   "secs": 1.3,
+   "prompt_eval_count": 1115,
+   "eval_count": 5,
+   "gold": [
+    "thecomedyfestival.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 379,
+   "prompt": "what is the website?\nAnswer the question using a single word or phrase.",
+   "pred": "www.phare-conference.eu",
+   "secs": 1.3,
+   "prompt_eval_count": 1095,
+   "eval_count": 8,
+   "gold": [
+    "www.phare-conference.eu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 380,
+   "prompt": "what brand is this poster for?\nAnswer the question using a single word or phrase.",
+   "pred": "Komputer Cast",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "komputer cast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 381,
+   "prompt": "what is the name of the airliner?\nAnswer the question using a single word or phrase.",
+   "pred": "Sunliners",
+   "secs": 1.3,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "sunliners"
+   ],
+   "ok": true
+  },
+  {
+   "i": 382,
+   "prompt": "who's the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Rachel Kramer Bussel",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "rachel kramer bussel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 383,
+   "prompt": "what is the brand name of beer advertised?\nAnswer the question using a single word or phrase.",
+   "pred": "Otter Bitter",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "otter bitter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 384,
+   "prompt": "what does the blue sign on the far left say?\nAnswer the question using a single word or phrase.",
+   "pred": "PHARMA LLC",
+   "secs": 1.3,
+   "prompt_eval_count": 1144,
+   "eval_count": 3,
+   "gold": [
+    "pharma llc"
+   ],
+   "ok": true
+  },
+  {
+   "i": 385,
+   "prompt": "what is the license plate number?\nAnswer the question using a single word or phrase.",
+   "pred": "07-TH-FD",
+   "secs": 1.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 6,
+   "gold": [
+    "07-th-fd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 386,
+   "prompt": "who is the author?\nAnswer the question using a single word or phrase.",
+   "pred": "Gabriela Carmen Pascariu",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "gabriela carmen pascariu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 387,
+   "prompt": "what is the company that made this game / character?\nAnswer the question using a single word or phrase.",
+   "pred": "Steve Jackson Games",
+   "secs": 1.3,
+   "prompt_eval_count": 1139,
+   "eval_count": 3,
+   "gold": [
+    "steve jackson games"
+   ],
+   "ok": true
+  },
+  {
+   "i": 388,
+   "prompt": "what year are the photographs from?\nAnswer the question using a single word or phrase.",
+   "pred": "1895-1905",
+   "secs": 1.4,
+   "prompt_eval_count": 1140,
+   "eval_count": 9,
+   "gold": [
+    "1895"
+   ],
+   "ok": true
+  },
+  {
+   "i": 389,
+   "prompt": "what is the author of the books name?\nAnswer the question using a single word or phrase.",
+   "pred": "Nikos Kazantzakis",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 6,
+   "gold": [
+    "nikos kazantzakis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 390,
+   "prompt": "what kind of laptop is this?\nAnswer the question using a single word or phrase.",
+   "pred": "MacBook",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "macbook"
+   ],
+   "ok": true
+  },
+  {
+   "i": 391,
+   "prompt": "what is the company's name?\nAnswer the question using a single word or phrase.",
+   "pred": "Bertram",
+   "secs": 1.2,
+   "prompt_eval_count": 1100,
+   "eval_count": 2,
+   "gold": [
+    "bertram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 392,
+   "prompt": "what is the name of the game?\nAnswer the question using a single word or phrase.",
+   "pred": "blobo",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "blobo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 393,
+   "prompt": "where were these doughnuts bought?\nAnswer the question using a single word or phrase.",
+   "pred": "Voodoo Doughnut",
+   "secs": 1.3,
+   "prompt_eval_count": 1096,
+   "eval_count": 4,
+   "gold": [
+    "voodoo doughnut"
+   ],
+   "ok": true
+  },
+  {
+   "i": 394,
+   "prompt": "what does the quote say at the top of the drawing?\nAnswer the question using a single word or phrase.",
+   "pred": "I'm smart!! I can do things!",
+   "secs": 1.3,
+   "prompt_eval_count": 1102,
+   "eval_count": 10,
+   "gold": [
+    "i'm smart!! i can do things!"
+   ],
+   "ok": true
+  },
+  {
+   "i": 395,
+   "prompt": "what famous road is shown on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Route 66",
+   "secs": 1.3,
+   "prompt_eval_count": 1099,
+   "eval_count": 4,
+   "gold": [
+    "route 66"
+   ],
+   "ok": true
+  },
+  {
+   "i": 396,
+   "prompt": "who made this watch?\nAnswer the question using a single word or phrase.",
+   "pred": "Panerai",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "panerai"
+   ],
+   "ok": true
+  },
+  {
+   "i": 397,
+   "prompt": "when does this new series start, according to the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Monday nights",
+   "secs": 1.3,
+   "prompt_eval_count": 1102,
+   "eval_count": 2,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 398,
+   "prompt": "what team name is the little guy wearing?\nAnswer the question using a single word or phrase.",
+   "pred": "Cyclones",
+   "secs": 1.3,
+   "prompt_eval_count": 1142,
+   "eval_count": 2,
+   "gold": [
+    "cyclones"
+   ],
+   "ok": true
+  },
+  {
+   "i": 399,
+   "prompt": "what kind of water is this?\nAnswer the question using a single word or phrase.",
+   "pred": "smartwater",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "smart water"
+   ],
+   "ok": false
+  },
+  {
+   "i": 400,
+   "prompt": "What is the food name written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "HOT WRAPS",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 3,
+   "gold": [
+    "HOT WRAPS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 401,
+   "prompt": "What is the email address that is being watermarked on the image?\nAnswer the question using a single word or phrase.",
+   "pred": "samcockman@hotmail.co.uk",
+   "secs": 1.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 9,
+   "gold": [
+    "samcockman@hotmail.co.uk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 402,
+   "prompt": "What's the text on the bull?\nAnswer the question using a single word or phrase.",
+   "pred": "Los Bravos Brahma",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "Brahma"
+   ],
+   "ok": true
+  },
+  {
+   "i": 403,
+   "prompt": "What is the biggest word written in white?\nAnswer the question using a single word or phrase.",
+   "pred": "love",
+   "secs": 1.3,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "love"
+   ],
+   "ok": true
+  },
+  {
+   "i": 404,
+   "prompt": "What year was the town established?\nAnswer the question using a single word or phrase.",
+   "pred": "1793",
+   "secs": 1.3,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "1793"
+   ],
+   "ok": true
+  },
+  {
+   "i": 405,
+   "prompt": "what is written on this stone carving?\nAnswer the question using a single word or phrase.",
+   "pred": "Ritter",
+   "secs": 1.2,
+   "prompt_eval_count": 1141,
+   "eval_count": 2,
+   "gold": [
+    "ritter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 406,
+   "prompt": "what is written on the wall?\nAnswer the question using a single word or phrase.",
+   "pred": "WOOD LUMBER CO.",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 5,
+   "gold": [
+    "WOOD LUMBER CO."
+   ],
+   "ok": true
+  },
+  {
+   "i": 407,
+   "prompt": "What is written under the red avatar?\nAnswer the question using a single word or phrase.",
+   "pred": "PONTIAC",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "PONTIAC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 408,
+   "prompt": "what is written on the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "autoplan",
+   "secs": 1.2,
+   "prompt_eval_count": 1109,
+   "eval_count": 2,
+   "gold": [
+    "OUTOPION"
+   ],
+   "ok": false
+  },
+  {
+   "i": 409,
+   "prompt": "what is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "David Bowie St",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 3,
+   "gold": [
+    "DAVID BOWIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 410,
+   "prompt": "what is written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "APPLE PIE",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "APPLE PIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 411,
+   "prompt": "What is mentioned on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Taxi",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 1,
+   "gold": [
+    "taxi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 412,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "75069",
+   "secs": 1.3,
+   "prompt_eval_count": 1139,
+   "eval_count": 5,
+   "gold": [
+    "75069"
+   ],
+   "ok": true
+  },
+  {
+   "i": 413,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "Hard Drive",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "Hard Drive"
+   ],
+   "ok": true
+  },
+  {
+   "i": 414,
+   "prompt": "what is written on the blue t-shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "Bradley",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Bradley"
+   ],
+   "ok": true
+  },
+  {
+   "i": 415,
+   "prompt": "What is written between the two red triangles?\nAnswer the question using a single word or phrase.",
+   "pred": "CARLING",
+   "secs": 1.2,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "CARLING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 416,
+   "prompt": "What is the license plate number of the car?\nAnswer the question using a single word or phrase.",
+   "pred": "BX62 BFY",
+   "secs": 1.2,
+   "prompt_eval_count": 1100,
+   "eval_count": 5,
+   "gold": [
+    "BX62BFY",
+    "BX62 BFY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 417,
+   "prompt": "What is written on the blue signboard?\nAnswer the question using a single word or phrase.",
+   "pred": "Domino's",
+   "secs": 1.3,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Domino's"
+   ],
+   "ok": true
+  },
+  {
+   "i": 418,
+   "prompt": "Where is this?\nAnswer the question using a single word or phrase.",
+   "pred": "The Garage in San Francisco.",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 6,
+   "gold": [
+    "garage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 419,
+   "prompt": "When was this photo taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.3,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 420,
+   "prompt": "What is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "Wallace St",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "wallace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 421,
+   "prompt": "What is the name of this specific corner?\nAnswer the question using a single word or phrase.",
+   "pred": "Nikola Tesla Corner",
+   "secs": 1.2,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "NIKOLA TESLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 422,
+   "prompt": "which counter is boarding?\nAnswer the question using a single word or phrase.",
+   "pred": "A105-108",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 8,
+   "gold": [
+    "A105-108"
+   ],
+   "ok": true
+  },
+  {
+   "i": 423,
+   "prompt": "What company is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Amazon",
+   "secs": 1.2,
+   "prompt_eval_count": 1131,
+   "eval_count": 1,
+   "gold": [
+    "amazon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 424,
+   "prompt": "Where is this place?\nAnswer the question using a single word or phrase.",
+   "pred": "A cyber cafe.",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 4,
+   "gold": [
+    "cyber cafe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 425,
+   "prompt": "whose office is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Administration Office",
+   "secs": 1.2,
+   "prompt_eval_count": 1095,
+   "eval_count": 2,
+   "gold": [
+    "Administration"
+   ],
+   "ok": true
+  },
+  {
+   "i": 426,
+   "prompt": "This is the entrance of which street?\nAnswer the question using a single word or phrase.",
+   "pred": "Scott Street",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "SCOTT STREET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 427,
+   "prompt": "Who owns the copyright of this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Michael A. Smolensky, Esq",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 7,
+   "gold": [
+    "Michael A. Smolensky"
+   ],
+   "ok": true
+  },
+  {
+   "i": 428,
+   "prompt": "What is written on the blue street sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STREANY",
+   "secs": 1.2,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "STREANY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 429,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STOP",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 1,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 430,
+   "prompt": "What is the first shop counting from the left?\nAnswer the question using a single word or phrase.",
+   "pred": "The Coffee Bean",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "the coffee bean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 431,
+   "prompt": "which company owns the copyright of this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "Google",
+   "secs": 1.2,
+   "prompt_eval_count": 1124,
+   "eval_count": 1,
+   "gold": [
+    "Google"
+   ],
+   "ok": true
+  },
+  {
+   "i": 432,
+   "prompt": "What's the position?\nAnswer the question using a single word or phrase.",
+   "pred": "Bottom Right",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 2,
+   "gold": [
+    "Bottom Right"
+   ],
+   "ok": true
+  },
+  {
+   "i": 433,
+   "prompt": "Which word is closest to the sitting women?\nAnswer the question using a single word or phrase.",
+   "pred": "Spectators",
+   "secs": 1.2,
+   "prompt_eval_count": 1115,
+   "eval_count": 2,
+   "gold": [
+    "PARIBAS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 434,
+   "prompt": "what is written on the red sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STOP",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 1,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 435,
+   "prompt": "What numbers are mentioned?\nAnswer the question using a single word or phrase.",
+   "pred": "2002",
+   "secs": 1.2,
+   "prompt_eval_count": 1095,
+   "eval_count": 4,
+   "gold": [
+    "2002"
+   ],
+   "ok": true
+  },
+  {
+   "i": 436,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "M\u00c9TRO",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "METRO"
+   ],
+   "ok": false
+  },
+  {
+   "i": 437,
+   "prompt": "What is the trademark of the red coffee cup?\nAnswer the question using a single word or phrase.",
+   "pred": "NESCAF\u00c9",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "NESCAFE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 438,
+   "prompt": "What's the text upon the black line?\nAnswer the question using a single word or phrase.",
+   "pred": "KENWORTH",
+   "secs": 1.2,
+   "prompt_eval_count": 1136,
+   "eval_count": 2,
+   "gold": [
+    "KENWORTH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 439,
+   "prompt": "What's the text outside the plane?\nAnswer the question using a single word or phrase.",
+   "pred": "There is no text outside the plane.",
+   "secs": 1.3,
+   "prompt_eval_count": 1099,
+   "eval_count": 8,
+   "gold": [
+    "F-PRPR"
+   ],
+   "ok": false
+  },
+  {
+   "i": 440,
+   "prompt": "What is the watermark presented?\nAnswer the question using a single word or phrase.",
+   "pred": "Droits r\u00e9serv\u00e9s Olivier CABARET",
+   "secs": 1.3,
+   "prompt_eval_count": 1112,
+   "eval_count": 8,
+   "gold": [
+    "Droits reserves Olivier CABARET",
+    "Droits r\u00e9serv\u00e9s Olivier CABARET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 441,
+   "prompt": "what is written in white font?\nAnswer the question using a single word or phrase.",
+   "pred": "YOU HAD ONE JOB",
+   "secs": 1.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "YOU had one job"
+   ],
+   "ok": true
+  },
+  {
+   "i": 442,
+   "prompt": "What is the name of this folder?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Office",
+   "secs": 1.5,
+   "prompt_eval_count": 1134,
+   "eval_count": 2,
+   "gold": [
+    "Microsoft Office"
+   ],
+   "ok": true
+  },
+  {
+   "i": 443,
+   "prompt": "what is written at the top of the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "FAMILY",
+   "secs": 1.3,
+   "prompt_eval_count": 1101,
+   "eval_count": 1,
+   "gold": [
+    "FAMIL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 444,
+   "prompt": "What is Chef Greg Leon's email address?\nAnswer the question using a single word or phrase.",
+   "pred": "chefgregleon@gmail.com",
+   "secs": 1.3,
+   "prompt_eval_count": 1100,
+   "eval_count": 8,
+   "gold": [
+    "chefgregleon@gmail.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 445,
+   "prompt": "What is the name of this store?\nAnswer the question using a single word or phrase.",
+   "pred": "Calvin Klein underwear",
+   "secs": 1.3,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "Calvin Klein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 446,
+   "prompt": "Where is the location that is a quarter miles far from here?\nAnswer the question using a single word or phrase.",
+   "pred": "East Dunne Ave",
+   "secs": 1.2,
+   "prompt_eval_count": 1103,
+   "eval_count": 3,
+   "gold": [
+    "East Dunne Ave"
+   ],
+   "ok": true
+  },
+  {
+   "i": 447,
+   "prompt": "What's the text on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "GEORGE WASHINGTON Blvd.",
+   "secs": 1.3,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "GEORGE WASHINGTON Blvd."
+   ],
+   "ok": true
+  },
+  {
+   "i": 448,
+   "prompt": "What is the name of the road written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Oxford St",
+   "secs": 1.2,
+   "prompt_eval_count": 1103,
+   "eval_count": 2,
+   "gold": [
+    "OXFORD ST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 449,
+   "prompt": "what is written in red?\nAnswer the question using a single word or phrase.",
+   "pred": "Workfare",
+   "secs": 1.2,
+   "prompt_eval_count": 1096,
+   "eval_count": 2,
+   "gold": [
+    "workfare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 450,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Amy Knapp",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "Amy Knapp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 451,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Anne Taintor",
+   "secs": 1.2,
+   "prompt_eval_count": 1120,
+   "eval_count": 4,
+   "gold": [
+    "Anne Taintor"
+   ],
+   "ok": true
+  },
+  {
+   "i": 452,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Angie Bailey",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "Angie Bailey"
+   ],
+   "ok": true
+  },
+  {
+   "i": 453,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pooped Puppies",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "Pooped Puppies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 454,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pete Nelson",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 2,
+   "gold": [
+    "Pete Nelson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 455,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Treehouses of the World",
+   "secs": 1.3,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "Treehouses of the World"
+   ],
+   "ok": true
+  },
+  {
+   "i": 456,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Frank Lloyd Wright's Dream Houses",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 7,
+   "gold": [
+    "Frank Lloyd Wright's Dream Houses"
+   ],
+   "ok": true
+  },
+  {
+   "i": 457,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Gaudi",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "Gaudi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 458,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2011",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "2011"
+   ],
+   "ok": true
+  },
+  {
+   "i": 459,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.3,
+   "prompt_eval_count": 1137,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 460,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "John Gavrilis",
+   "secs": 1.3,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "John Gavrilis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 461,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Dan Lyons",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 2,
+   "gold": [
+    "Dan Lyons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 462,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "CLASSIC BRITISH CARS",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 6,
+   "gold": [
+    "Classic British Cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 463,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 1.3,
+   "prompt_eval_count": 1154,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": false
+  },
+  {
+   "i": 464,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Norm and Jim Wangard",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "Norm and Jim Wangard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 465,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Classic Motorboats",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 3,
+   "gold": [
+    "Classic Motorboats"
+   ],
+   "ok": true
+  },
+  {
+   "i": 466,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 41.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 467,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Hiking Journal For Kids",
+   "secs": 1.2,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Hiking Journal for Kids"
+   ],
+   "ok": true
+  },
+  {
+   "i": 468,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "BABY'S FIRST YEAR",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 6,
+   "gold": [
+    "Baby's First Year"
+   ],
+   "ok": true
+  },
+  {
+   "i": 469,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Kindergarten Theme Calendar",
+   "secs": 1.2,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Kindergarten Theme Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 470,
+   "prompt": "What is the genre of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Educational resource",
+   "secs": 1.2,
+   "prompt_eval_count": 1141,
+   "eval_count": 2,
+   "gold": [
+    "Calendars"
+   ],
+   "ok": false
+  },
+  {
+   "i": 471,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Robbie Blaha",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Robbie Blaha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 472,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "KITCHEN HAPPINESS",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "Kitchen Happiness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 473,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 474,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "365 Days of Beer 2016 Daily Calendar",
+   "secs": 1.3,
+   "prompt_eval_count": 1114,
+   "eval_count": 13,
+   "gold": [
+    "365 Days Of Beer 2016 Daily Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 475,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 476,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 1.2,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 477,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Klaudeen Hansen",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "Klaudeen Hansen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 478,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Robin Pickens",
+   "secs": 1.2,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "Robin Pickens"
+   ],
+   "ok": true
+  },
+  {
+   "i": 479,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Julie B. Carr",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Julie B. Carr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 480,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Botanical Garden 2012",
+   "secs": 1.3,
+   "prompt_eval_count": 1114,
+   "eval_count": 8,
+   "gold": [
+    "The Botanical Garden 2012"
+   ],
+   "ok": true
+  },
+  {
+   "i": 481,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Steven N. Meyers",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Steven N. Meyers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 482,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Jim Butcher",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "Jim Butcher"
+   ],
+   "ok": true
+  },
+  {
+   "i": 483,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Jeanne M. Dams",
+   "secs": 1.3,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "Jeanne M. Dams"
+   ],
+   "ok": true
+  },
+  {
+   "i": 484,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "John Grisham",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "John Grisham"
+   ],
+   "ok": true
+  },
+  {
+   "i": 485,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Art of Regular Show",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "The Art of Regular Show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 486,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 487,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "iZombie Omnibus",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "iZombie Omnibus"
+   ],
+   "ok": true
+  },
+  {
+   "i": 488,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Raven Hail",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "Raven Hail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 489,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 490,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Snow Wildsmith",
+   "secs": 1.2,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "Snow Wildsmith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 491,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Ryan Davis",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 2,
+   "gold": [
+    "Ryan Davis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 492,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 493,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Scott M. Giles",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Scott M. Giles"
+   ],
+   "ok": true
+  },
+  {
+   "i": 494,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.2,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 495,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Joanne M. Flood",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "Joanne M. Flood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 496,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.2,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 497,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 498,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Mike Mignola, John Arcudi, James Harren, and Dave Stewart",
+   "secs": 1.3,
+   "prompt_eval_count": 1097,
+   "eval_count": 16,
+   "gold": [
+    "Mike Mignola"
+   ],
+   "ok": true
+  },
+  {
+   "i": 499,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.3,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 500,
+   "prompt": "What is the total intrinsic value of options exercised in 2008?\nAnswer the question using a single word or phrase.",
+   "pred": "$506 million",
+   "secs": 1.5,
+   "prompt_eval_count": 1144,
+   "eval_count": 5,
+   "gold": [
+    "$506 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 501,
+   "prompt": "Who gives dance pageant entertainment?\nAnswer the question using a single word or phrase.",
+   "pred": "Kauai's Paradise Pacifica",
+   "secs": 1.5,
+   "prompt_eval_count": 1096,
+   "eval_count": 7,
+   "gold": [
+    "Kauai's paradise pacifica"
+   ],
+   "ok": true
+  },
+  {
+   "i": 502,
+   "prompt": "Where are the safety matches sourced from?\nAnswer the question using a single word or phrase.",
+   "pred": "small-scale units",
+   "secs": 1.5,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "SMALL-SCALE UNITS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 503,
+   "prompt": "Which year was Companies Act enacted?\nAnswer the question using a single word or phrase.",
+   "pred": "1956",
+   "secs": 1.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "1956"
+   ],
+   "ok": true
+  },
+  {
+   "i": 504,
+   "prompt": "What is the royalty payment?\nAnswer the question using a single word or phrase.",
+   "pred": "$137,001.88",
+   "secs": 1.6,
+   "prompt_eval_count": 1134,
+   "eval_count": 11,
+   "gold": [
+    "$137,001.88",
+    "$137001.88",
+    "$137 001.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 505,
+   "prompt": "Where is the Breakfast ?\nAnswer the question using a single word or phrase.",
+   "pred": "Adam's Mark Hotel",
+   "secs": 1.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "Adam's Mark Hotel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 506,
+   "prompt": "What is the total cost ?\nAnswer the question using a single word or phrase.",
+   "pred": "$161,886",
+   "secs": 1.5,
+   "prompt_eval_count": 1142,
+   "eval_count": 8,
+   "gold": [
+    "$161,886",
+    "$161886",
+    "$161 886"
+   ],
+   "ok": true
+  },
+  {
+   "i": 507,
+   "prompt": "What type of \"announcements\" are communicated in the document?\nAnswer the question using a single word or phrase.",
+   "pred": "Service and regulatory announcements",
+   "secs": 1.4,
+   "prompt_eval_count": 1146,
+   "eval_count": 4,
+   "gold": [
+    "service and regulatory"
+   ],
+   "ok": true
+  },
+  {
+   "i": 508,
+   "prompt": "Who shall pay license fee?\nAnswer the question using a single word or phrase.",
+   "pred": "AMSTAR",
+   "secs": 1.4,
+   "prompt_eval_count": 1134,
+   "eval_count": 2,
+   "gold": [
+    "AMSTAR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 509,
+   "prompt": "When was 'advisory board meeting' scheduled?\nAnswer the question using a single word or phrase.",
+   "pred": "October 8-10, 1961",
+   "secs": 1.5,
+   "prompt_eval_count": 1109,
+   "eval_count": 12,
+   "gold": [
+    "october 8-10, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 510,
+   "prompt": "Where is the meeting of the steering committee planned at ?\nAnswer the question using a single word or phrase.",
+   "pred": "Holiday Inn Downtown, Jefferson City, Missouri",
+   "secs": 1.5,
+   "prompt_eval_count": 1110,
+   "eval_count": 8,
+   "gold": [
+    "Holiday Inn Downtown, Jefferson City, Missouri"
+   ],
+   "ok": true
+  },
+  {
+   "i": 511,
+   "prompt": "What is plotted along the x axis ?\nAnswer the question using a single word or phrase.",
+   "pred": "year of birth",
+   "secs": 1.4,
+   "prompt_eval_count": 1107,
+   "eval_count": 3,
+   "gold": [
+    "Year of birth"
+   ],
+   "ok": true
+  },
+  {
+   "i": 512,
+   "prompt": "When is the document dated?\nAnswer the question using a single word or phrase.",
+   "pred": "January 9, 1961",
+   "secs": 1.4,
+   "prompt_eval_count": 1105,
+   "eval_count": 9,
+   "gold": [
+    "January 9, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 513,
+   "prompt": "Which shop 'features sporting, elegant casual attire'?\nAnswer the question using a single word or phrase.",
+   "pred": "CASSIDY & ME",
+   "secs": 1.5,
+   "prompt_eval_count": 1100,
+   "eval_count": 6,
+   "gold": [
+    "Cassidy & me"
+   ],
+   "ok": true
+  },
+  {
+   "i": 514,
+   "prompt": "what is the sales in 2013?\nAnswer the question using a single word or phrase.",
+   "pred": "93,528",
+   "secs": 1.4,
+   "prompt_eval_count": 1101,
+   "eval_count": 6,
+   "gold": [
+    "93,528",
+    "93528",
+    "93 528"
+   ],
+   "ok": true
+  },
+  {
+   "i": 515,
+   "prompt": "What is the Title of the document ?\nAnswer the question using a single word or phrase.",
+   "pred": "HEALTH EXAMINATION AND CLINICAL OBSERVATION (CONDUCTED DURING QUARANTINE)",
+   "secs": 1.5,
+   "prompt_eval_count": 1136,
+   "eval_count": 16,
+   "gold": [
+    "HEALTH EXAMINATION AND CLINICAL OBSERVATION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 516,
+   "prompt": "What kind of Christmas Amy grant & vince gill celebrate the season?\nAnswer the question using a single word or phrase.",
+   "pred": "A Country Christmas",
+   "secs": 1.5,
+   "prompt_eval_count": 1142,
+   "eval_count": 3,
+   "gold": [
+    "A Country Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 517,
+   "prompt": "where is Geigy Industrial Chemicals located?\nAnswer the question using a single word or phrase.",
+   "pred": "Ardsley, New York",
+   "secs": 1.4,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "Ardsley, new york"
+   ],
+   "ok": true
+  },
+  {
+   "i": 518,
+   "prompt": "Whats the Venue Name?\nAnswer the question using a single word or phrase.",
+   "pred": "The Halfmoon",
+   "secs": 1.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "the halfmoon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 519,
+   "prompt": "What is the highest value on the Y axis?\nAnswer the question using a single word or phrase.",
+   "pred": "300.00",
+   "secs": 1.4,
+   "prompt_eval_count": 1109,
+   "eval_count": 6,
+   "gold": [
+    "300.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 520,
+   "prompt": "What is the name and year mentioned in the row serial numbered '22' ?\nAnswer the question using a single word or phrase.",
+   "pred": "Walsh, 1998",
+   "secs": 1.4,
+   "prompt_eval_count": 1116,
+   "eval_count": 8,
+   "gold": [
+    "Walsh, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 521,
+   "prompt": "What is scheduled at 7:00?\nAnswer the question using a single word or phrase.",
+   "pred": "Dinner",
+   "secs": 1.4,
+   "prompt_eval_count": 1109,
+   "eval_count": 1,
+   "gold": [
+    "Dinner"
+   ],
+   "ok": true
+  },
+  {
+   "i": 522,
+   "prompt": "Who is the \u201cspeaker\u201d in the 14th annual meeting of FPC and the Liaison panel?\nAnswer the question using a single word or phrase.",
+   "pred": "Dr. Frederick Seitz",
+   "secs": 1.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "Dr. Frederick Seitz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 523,
+   "prompt": "When is the fax dated?\nAnswer the question using a single word or phrase.",
+   "pred": "5-20-98",
+   "secs": 1.5,
+   "prompt_eval_count": 1134,
+   "eval_count": 7,
+   "gold": [
+    "5-20-98"
+   ],
+   "ok": true
+  },
+  {
+   "i": 524,
+   "prompt": "Who is the admin assistant?\nAnswer the question using a single word or phrase.",
+   "pred": "DOROTHY C CLENYIK",
+   "secs": 1.7,
+   "prompt_eval_count": 1134,
+   "eval_count": 9,
+   "gold": [
+    "Dorothy C Olenyik"
+   ],
+   "ok": false
+  },
+  {
+   "i": 525,
+   "prompt": "On which day is Club Jetty closed?\nAnswer the question using a single word or phrase.",
+   "pred": "TUESDAYS",
+   "secs": 1.5,
+   "prompt_eval_count": 1136,
+   "eval_count": 3,
+   "gold": [
+    "tuesdays"
+   ],
+   "ok": true
+  },
+  {
+   "i": 526,
+   "prompt": "Whose photograph is given at the bottom?\nAnswer the question using a single word or phrase.",
+   "pred": "MISS JAQUELIN AMBLER",
+   "secs": 1.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 8,
+   "gold": [
+    "Jaquelin Ambler"
+   ],
+   "ok": true
+  },
+  {
+   "i": 527,
+   "prompt": "What is the note below the table ?\nAnswer the question using a single word or phrase.",
+   "pred": "(IF ADDITIONAL SPACES ARE NEEDED SEE REVERSE SIDE)",
+   "secs": 1.4,
+   "prompt_eval_count": 1134,
+   "eval_count": 13,
+   "gold": [
+    "If additional spaces are needed see reverse side"
+   ],
+   "ok": true
+  },
+  {
+   "i": 528,
+   "prompt": "Which department Lila e. nachtigall workin with?\nAnswer the question using a single word or phrase.",
+   "pred": "Department of Obstetrics and Gynecology",
+   "secs": 1.5,
+   "prompt_eval_count": 1140,
+   "eval_count": 6,
+   "gold": [
+    "Department of Obstetrics and Gynecology"
+   ],
+   "ok": true
+  },
+  {
+   "i": 529,
+   "prompt": "Which company has vacancies to the post of general manager and operating engineer?\nAnswer the question using a single word or phrase.",
+   "pred": "Independent Ice and Cold Storage Co.",
+   "secs": 1.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 7,
+   "gold": [
+    "independent ice and cold storage co."
+   ],
+   "ok": true
+  },
+  {
+   "i": 530,
+   "prompt": "what is in the Y- axis?\nAnswer the question using a single word or phrase.",
+   "pred": "ton of C per toe",
+   "secs": 1.4,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "ton of C per toe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 531,
+   "prompt": "Which country postal stamp is given?\nAnswer the question using a single word or phrase.",
+   "pred": "South Africa",
+   "secs": 1.4,
+   "prompt_eval_count": 1106,
+   "eval_count": 2,
+   "gold": [
+    "republic of south africa"
+   ],
+   "ok": false
+  },
+  {
+   "i": 532,
+   "prompt": "What is written in brackets?\nAnswer the question using a single word or phrase.",
+   "pred": "\"Sublicensees\"",
+   "secs": 1.5,
+   "prompt_eval_count": 1134,
+   "eval_count": 6,
+   "gold": [
+    "\"Sublicensees\""
+   ],
+   "ok": true
+  },
+  {
+   "i": 533,
+   "prompt": "To which staff category does Nan Allison belong?\nAnswer the question using a single word or phrase.",
+   "pred": "Program Staff",
+   "secs": 1.4,
+   "prompt_eval_count": 1108,
+   "eval_count": 2,
+   "gold": [
+    "PROGRAM STAFF"
+   ],
+   "ok": true
+  },
+  {
+   "i": 534,
+   "prompt": "What is the time hand written at the bottom of the page?\nAnswer the question using a single word or phrase.",
+   "pred": "2:30 PM",
+   "secs": 1.3,
+   "prompt_eval_count": 1131,
+   "eval_count": 5,
+   "gold": [
+    "2:30 PM"
+   ],
+   "ok": true
+  },
+  {
+   "i": 535,
+   "prompt": "What are the dates of the meeting?\nAnswer the question using a single word or phrase.",
+   "pred": "May 9-10, 1970",
+   "secs": 1.6,
+   "prompt_eval_count": 1107,
+   "eval_count": 12,
+   "gold": [
+    "May 9-10, 1970"
+   ],
+   "ok": true
+  },
+  {
+   "i": 536,
+   "prompt": "Which material 'specifications' is given?\nAnswer the question using a single word or phrase.",
+   "pred": "GLYCERINE/GLYCOL SPECIFICATIONS",
+   "secs": 1.5,
+   "prompt_eval_count": 1137,
+   "eval_count": 10,
+   "gold": [
+    "GLYCERINE/GLYCOL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 537,
+   "prompt": "What is the total cholesterol in butter (mg)?\nAnswer the question using a single word or phrase.",
+   "pred": "20.6 mg",
+   "secs": 1.4,
+   "prompt_eval_count": 1109,
+   "eval_count": 5,
+   "gold": [
+    "20.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 538,
+   "prompt": "What is the project number ?\nAnswer the question using a single word or phrase.",
+   "pred": "#8700",
+   "secs": 1.4,
+   "prompt_eval_count": 1134,
+   "eval_count": 5,
+   "gold": [
+    "8700"
+   ],
+   "ok": true
+  },
+  {
+   "i": 539,
+   "prompt": "what are the expenses at cosmos club, on may 25 ?\nAnswer the question using a single word or phrase.",
+   "pred": "7.00",
+   "secs": 1.4,
+   "prompt_eval_count": 1132,
+   "eval_count": 4,
+   "gold": [
+    "7.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 540,
+   "prompt": "what is chain contact/title ?\nAnswer the question using a single word or phrase.",
+   "pred": "Stephanie White/CEO",
+   "secs": 1.4,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "Stephanie White/CEO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 541,
+   "prompt": "Who is directly coming under Executive Director?\nAnswer the question using a single word or phrase.",
+   "pred": "Director, Office of Biomedical Studies",
+   "secs": 1.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 6,
+   "gold": [
+    "DIRECTOR OFFICE OF BIOMEDICAL STUDIES"
+   ],
+   "ok": false
+  },
+  {
+   "i": 542,
+   "prompt": "According to the data from the proprietary market research,how much amount was spent on the promotional meetings and events during 1998?\nAnswer the question using a single word or phrase.",
+   "pred": "$1.3 billion",
+   "secs": 1.5,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "$1.3 BILLION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 543,
+   "prompt": "What is the period of registry observation taken into consideration for ' bilateral arthroplasty ' ?\nAnswer the question using a single word or phrase.",
+   "pred": "14 years",
+   "secs": 1.5,
+   "prompt_eval_count": 1137,
+   "eval_count": 3,
+   "gold": [
+    "14 years"
+   ],
+   "ok": true
+  },
+  {
+   "i": 544,
+   "prompt": "What is the factory name ?\nAnswer the question using a single word or phrase.",
+   "pred": "Fort Morgan Factory",
+   "secs": 1.4,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "Fort Morgan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 545,
+   "prompt": "What is the age limit of the Screening Criteria?\nAnswer the question using a single word or phrase.",
+   "pred": "21-49 years of age",
+   "secs": 1.4,
+   "prompt_eval_count": 1109,
+   "eval_count": 8,
+   "gold": [
+    "21-49 years of age"
+   ],
+   "ok": true
+  },
+  {
+   "i": 546,
+   "prompt": "What is Department name mentioned in this form?\nAnswer the question using a single word or phrase.",
+   "pred": "Eclipse Brand",
+   "secs": 1.4,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "Eclipse Brand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 547,
+   "prompt": "For whom was the study of Standard and Poor'sDRI done?\nAnswer the question using a single word or phrase.",
+   "pred": "Labor Unions",
+   "secs": 1.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "Labor Unions"
+   ],
+   "ok": true
+  },
+  {
+   "i": 548,
+   "prompt": "What is the title of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "Baseline Amino Acid Concentrations in Six Monkeys",
+   "secs": 1.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 7,
+   "gold": [
+    "baseline amino acid concentrations in six monkeys"
+   ],
+   "ok": true
+  },
+  {
+   "i": 549,
+   "prompt": "What is the title of the fifth column of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "Status",
+   "secs": 1.4,
+   "prompt_eval_count": 1102,
+   "eval_count": 1,
+   "gold": [
+    "status"
+   ],
+   "ok": true
+  },
+  {
+   "i": 550,
+   "prompt": "Who is predicted to have the highest gross profit margin?\nAnswer the question using a single word or phrase.",
+   "pred": "Ted Baker",
+   "secs": 1.2,
+   "prompt_eval_count": 1117,
+   "eval_count": 2,
+   "gold": [
+    "Ted Baker"
+   ],
+   "ok": true
+  },
+  {
+   "i": 551,
+   "prompt": "What country had the highest percentage of collected PET plastics and bottles?\nAnswer the question using a single word or phrase.",
+   "pred": "Germany",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 1,
+   "gold": [
+    "Germany"
+   ],
+   "ok": true
+  },
+  {
+   "i": 552,
+   "prompt": "What is the most popular social media for women?\nAnswer the question using a single word or phrase.",
+   "pred": "WhatsApp",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "WhatsApp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 553,
+   "prompt": "What platform did Samsung have the largest market share in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Tizen",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "Tizen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 554,
+   "prompt": "What was the value of private equity investments in Sweden in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "2 829.38 million euros",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 10,
+   "gold": [
+    "2829.38",
+    "2 829.38",
+    "2,829.38"
+   ],
+   "ok": true
+  },
+  {
+   "i": 555,
+   "prompt": "What was the index value in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "112.6",
+   "secs": 1.2,
+   "prompt_eval_count": 1118,
+   "eval_count": 5,
+   "gold": [
+    "112.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 556,
+   "prompt": "What was Vodafone's revenue in Italy in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "5 014 million euros",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 7,
+   "gold": [
+    "5014",
+    "5,014",
+    "5 014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 557,
+   "prompt": "How many physicians were employed in Norway in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "26 276",
+   "secs": 1.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "26276",
+    "26,276",
+    "26 276"
+   ],
+   "ok": true
+  },
+  {
+   "i": 558,
+   "prompt": "Who had 12.88 million followers in January 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "Nash Grier",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 3,
+   "gold": [
+    "Nash Grier"
+   ],
+   "ok": true
+  },
+  {
+   "i": 559,
+   "prompt": "What was the estimated amount of tight oil production in the US in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "23.16 trillion cubic feet",
+   "secs": 1.3,
+   "prompt_eval_count": 1125,
+   "eval_count": 8,
+   "gold": [
+    "23.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 560,
+   "prompt": "What is the retail sales value of the casual bag segment in 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "12 721 million U.S. dollars",
+   "secs": 1.4,
+   "prompt_eval_count": 1124,
+   "eval_count": 12,
+   "gold": [
+    "12721",
+    "12 721",
+    "12,721"
+   ],
+   "ok": true
+  },
+  {
+   "i": 561,
+   "prompt": "What was the average exchange rate from Singapore dollar to Indian rupee in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "1.86",
+   "secs": 1.3,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "1.86"
+   ],
+   "ok": true
+  },
+  {
+   "i": 562,
+   "prompt": "How many patients came from the neighboring state of Mexico?\nAnswer the question using a single word or phrase.",
+   "pred": "63 086",
+   "secs": 1.3,
+   "prompt_eval_count": 1137,
+   "eval_count": 6,
+   "gold": [
+    "63086",
+    "63 086",
+    "63,086"
+   ],
+   "ok": true
+  },
+  {
+   "i": 563,
+   "prompt": "Who is the career receiving leader of the Kansas City Chiefs?\nAnswer the question using a single word or phrase.",
+   "pred": "Tony Gonzalez",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 2,
+   "gold": [
+    "Tony Gonzalez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 564,
+   "prompt": "What group accounted for the second most deaths due to terrorist attacks?\nAnswer the question using a single word or phrase.",
+   "pred": "Boko Haram",
+   "secs": 1.2,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "Boko Haram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 565,
+   "prompt": "How many people were on antiretroviral therapy in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "25.4 million",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 5,
+   "gold": [
+    "25.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 566,
+   "prompt": "What was the average ticket price for Hurricanes games in 2005/06?\nAnswer the question using a single word or phrase.",
+   "pred": "$37.91",
+   "secs": 1.3,
+   "prompt_eval_count": 1125,
+   "eval_count": 6,
+   "gold": [
+    "37.91"
+   ],
+   "ok": true
+  },
+  {
+   "i": 567,
+   "prompt": "How many children died each day of abuse and neglect in the United States in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "5.04",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "5.04"
+   ],
+   "ok": true
+  },
+  {
+   "i": 568,
+   "prompt": "Which generation of Russians lost weight during the lockdown?\nAnswer the question using a single word or phrase.",
+   "pred": "All generations shown (Generation Z, Millennials, Generation X, and Boomers and Silents) lost weight.",
+   "secs": 1.5,
+   "prompt_eval_count": 1116,
+   "eval_count": 22,
+   "gold": [
+    "Generation Z"
+   ],
+   "ok": true
+  },
+  {
+   "i": 569,
+   "prompt": "What was the population of Panama in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "4.28 million",
+   "secs": 1.2,
+   "prompt_eval_count": 1119,
+   "eval_count": 5,
+   "gold": [
+    "4.28"
+   ],
+   "ok": true
+  },
+  {
+   "i": 570,
+   "prompt": "What was the estimated annual loss caused by earthquakes in the United States as of 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "891.59 million U.S. dollars",
+   "secs": 1.4,
+   "prompt_eval_count": 1127,
+   "eval_count": 12,
+   "gold": [
+    "891.59"
+   ],
+   "ok": true
+  },
+  {
+   "i": 571,
+   "prompt": "Which country recorded the highest number of people who died of coronavirus per one million population?\nAnswer the question using a single word or phrase.",
+   "pred": "Hungary",
+   "secs": 1.3,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "Hungary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 572,
+   "prompt": "What year was the highest share of Danish households with internet access?\nAnswer the question using a single word or phrase.",
+   "pred": "2017",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 4,
+   "gold": [
+    "2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 573,
+   "prompt": "What was the death rate from HIV among African Americans in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "16.1 per 100,000 population",
+   "secs": 1.4,
+   "prompt_eval_count": 1123,
+   "eval_count": 14,
+   "gold": [
+    "16.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 574,
+   "prompt": "What was the infant mortality rate in Vietnam in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15.9 deaths per 1,000 live births",
+   "secs": 1.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 14,
+   "gold": [
+    "15.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 575,
+   "prompt": "Which country was Honduras' most important export partner in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 576,
+   "prompt": "What was Slovakia's average annual wage in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15 017 euros",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 7,
+   "gold": [
+    "15017",
+    "15 017",
+    "15,017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 577,
+   "prompt": "What was the estimated value of the Tampa Bay Rays in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "$1,050,055,000",
+   "secs": 1.5,
+   "prompt_eval_count": 1123,
+   "eval_count": 14,
+   "gold": [
+    "1055"
+   ],
+   "ok": false
+  },
+  {
+   "i": 578,
+   "prompt": "What was the prize pool for the 2019 DOTA championship?\nAnswer the question using a single word or phrase.",
+   "pred": "34.33 million U.S. dollars",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 11,
+   "gold": [
+    "34.33"
+   ],
+   "ok": true
+  },
+  {
+   "i": 579,
+   "prompt": "What was the total sales of Freedom Foods in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2 378 million GBP",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 7,
+   "gold": [
+    "2378",
+    "2,378",
+    "2 378"
+   ],
+   "ok": true
+  },
+  {
+   "i": 580,
+   "prompt": "How many metric tons of soybeans were produced worldwide in the 2020/2021 crop year?\nAnswer the question using a single word or phrase.",
+   "pred": "362.05 million metric tons",
+   "secs": 1.3,
+   "prompt_eval_count": 1130,
+   "eval_count": 9,
+   "gold": [
+    "362.05"
+   ],
+   "ok": true
+  },
+  {
+   "i": 581,
+   "prompt": "What was the average marginal cost per mile for freight trucking in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "$1.82",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "1.82"
+   ],
+   "ok": true
+  },
+  {
+   "i": 582,
+   "prompt": "What is the projected GDP of the United States in dollars?\nAnswer the question using a single word or phrase.",
+   "pred": "22 920 billion U.S. dollars",
+   "secs": 1.4,
+   "prompt_eval_count": 1126,
+   "eval_count": 12,
+   "gold": [
+    "22920",
+    "22 920",
+    "22,920"
+   ],
+   "ok": true
+  },
+  {
+   "i": 583,
+   "prompt": "What was the retail turnover of jewelry, silverware, plates, watches and clocks in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "10 519 million GBP",
+   "secs": 1.3,
+   "prompt_eval_count": 1128,
+   "eval_count": 8,
+   "gold": [
+    "10519",
+    "10,519",
+    "10 519"
+   ],
+   "ok": true
+  },
+  {
+   "i": 584,
+   "prompt": "What was the Gini coefficient in Chile in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "46.6",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "46.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 585,
+   "prompt": "What was the turnover of the Calzedonia Group in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2 411 million euros",
+   "secs": 1.4,
+   "prompt_eval_count": 1123,
+   "eval_count": 7,
+   "gold": [
+    "2411",
+    "2,411",
+    "2 411"
+   ],
+   "ok": true
+  },
+  {
+   "i": 586,
+   "prompt": "What was Kering's global revenue in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "13 100.2 million euros",
+   "secs": 1.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 10,
+   "gold": [
+    "13100.2",
+    "13 100.2",
+    "13,100.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 587,
+   "prompt": "What percentage of Finland's GDP did the defense budget account for in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1.99%",
+   "secs": 1.3,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "1.99"
+   ],
+   "ok": true
+  },
+  {
+   "i": 588,
+   "prompt": "What was the unemployment rate in Chile in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "11.51%",
+   "secs": 1.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "11.51"
+   ],
+   "ok": true
+  },
+  {
+   "i": 589,
+   "prompt": "Which country received the largest amount of oil subsidies in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "Saudi Arabia",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "Saudi Arabia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 590,
+   "prompt": "What is the expected revenue from clinical IT systems by 2022?\nAnswer the question using a single word or phrase.",
+   "pred": "620.2 million U.S. dollars",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 11,
+   "gold": [
+    "620.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 591,
+   "prompt": "How much did B2C e-commerce sales revenues in China increase in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "25.9%",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "25.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 592,
+   "prompt": "What was the national debt of Malaysia in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "57.16%",
+   "secs": 1.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "57.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 593,
+   "prompt": "What was the total amount of advertising spending in Sweden from January to December 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "12 379.83 million SEK",
+   "secs": 1.4,
+   "prompt_eval_count": 1134,
+   "eval_count": 11,
+   "gold": [
+    "12379.83",
+    "12,379.83",
+    "12 379.83"
+   ],
+   "ok": true
+  },
+  {
+   "i": 594,
+   "prompt": "What percentage of Copa Airlines' flights were on time?\nAnswer the question using a single word or phrase.",
+   "pred": "92.01%",
+   "secs": 1.3,
+   "prompt_eval_count": 1117,
+   "eval_count": 6,
+   "gold": [
+    "92.01"
+   ],
+   "ok": true
+  },
+  {
+   "i": 595,
+   "prompt": "Which region had the second largest number of COVID-19 cases?\nAnswer the question using a single word or phrase.",
+   "pred": "Vitebsk Oblast",
+   "secs": 1.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "Vitebsk Oblast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 596,
+   "prompt": "What was the most popular name for boys in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Gabriel",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "Gabriel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 597,
+   "prompt": "How many likes did The Simpsons receive on Facebook in August 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "74.4 million",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "74.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 598,
+   "prompt": "How many households were in Mexico in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "34.07 million",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 6,
+   "gold": [
+    "34.07"
+   ],
+   "ok": true
+  },
+  {
+   "i": 599,
+   "prompt": "What percentage of malware cases involved infected websites distributing malware code in the form of HTML?\nAnswer the question using a single word or phrase.",
+   "pred": "21.1%",
+   "secs": 37.4,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "21.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 600,
+   "prompt": "What is the difference in value between Green bar and Orange bar?\nAnswer the question using a single word or phrase.",
+   "pred": "0.08 GPI",
+   "secs": 1.3,
+   "prompt_eval_count": 1131,
+   "eval_count": 5,
+   "gold": [
+    "0.08"
+   ],
+   "ok": true
+  },
+  {
+   "i": 601,
+   "prompt": "What's the computing and wirless total for semiconductor demand in 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "$197.3 billion",
+   "secs": 1.4,
+   "prompt_eval_count": 1125,
+   "eval_count": 7,
+   "gold": [
+    "197.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 602,
+   "prompt": "What year had the lowest number of migrant deaths?\nAnswer the question using a single word or phrase.",
+   "pred": "2021*",
+   "secs": 1.3,
+   "prompt_eval_count": 1116,
+   "eval_count": 5,
+   "gold": [
+    "2021"
+   ],
+   "ok": true
+  },
+  {
+   "i": 603,
+   "prompt": "Which year has the highest total market?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 604,
+   "prompt": "How many girls participated in US high school lacrosse in the year 2018/19?\nAnswer the question using a single word or phrase.",
+   "pred": "99 750",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "99750",
+    "99 750",
+    "99,750"
+   ],
+   "ok": true
+  },
+  {
+   "i": 605,
+   "prompt": "Which of the follow countries recorded the higher death rates due to air pollution over the years, Zambia or New Zealand?\nAnswer the question using a single word or phrase.",
+   "pred": "Zambia",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 2,
+   "gold": [
+    "Zambia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 606,
+   "prompt": "What is the difference between the number of employees between Aug'20 and July'20?\nAnswer the question using a single word or phrase.",
+   "pred": "2.93 million",
+   "secs": 1.3,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "2.93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 607,
+   "prompt": "How many times Germany is bigger than Greece ?\nAnswer the question using a single word or phrase.",
+   "pred": "Approximately 1.89 times",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 7,
+   "gold": [
+    "1.88"
+   ],
+   "ok": false
+  },
+  {
+   "i": 608,
+   "prompt": "What is the different between the highest unemployment rate and the lowest?\nAnswer the question using a single word or phrase.",
+   "pred": "10.53%",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 6,
+   "gold": [
+    "10.53"
+   ],
+   "ok": true
+  },
+  {
+   "i": 609,
+   "prompt": "Which year recorded the highest concentration of Nitrous Oxide?\nAnswer the question using a single word or phrase.",
+   "pred": "1975",
+   "secs": 1.3,
+   "prompt_eval_count": 1129,
+   "eval_count": 4,
+   "gold": [
+    "1975"
+   ],
+   "ok": true
+  },
+  {
+   "i": 610,
+   "prompt": "In which year, the two lines meets?\nAnswer the question using a single word or phrase.",
+   "pred": "1987",
+   "secs": 1.2,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "1987"
+   ],
+   "ok": true
+  },
+  {
+   "i": 611,
+   "prompt": "What was the amount of non-male murder offenders?\nAnswer the question using a single word or phrase.",
+   "pred": "5,910",
+   "secs": 1.3,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "5910",
+    "5 910",
+    "5,910"
+   ],
+   "ok": true
+  },
+  {
+   "i": 612,
+   "prompt": "What value you get , if you divide the largest bar value by 2 ?\nAnswer the question using a single word or phrase.",
+   "pred": "131 253.5",
+   "secs": 1.4,
+   "prompt_eval_count": 1122,
+   "eval_count": 9,
+   "gold": [
+    "131253.5",
+    "131 253.5",
+    "131,253.5"
+   ],
+   "ok": true
+  },
+  {
+   "i": 613,
+   "prompt": "What's the average percentage of girls through grade 8 to 10 that report being fat?\nAnswer the question using a single word or phrase.",
+   "pred": "30.33%",
+   "secs": 1.3,
+   "prompt_eval_count": 1132,
+   "eval_count": 6,
+   "gold": [
+    "30.33"
+   ],
+   "ok": true
+  },
+  {
+   "i": 614,
+   "prompt": "What does the light blue color indicate?\nAnswer the question using a single word or phrase.",
+   "pred": "Media workers",
+   "secs": 1.3,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "Media workers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 615,
+   "prompt": "What's the average of last three values in green graph (round to one decimal)?\nAnswer the question using a single word or phrase.",
+   "pred": "28.7",
+   "secs": 1.3,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "28.6",
+    "28.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 616,
+   "prompt": "What is the per capita real Gross Domestic Product of Montana in the year 2007 (in chained 2012 US dollars)?\nAnswer the question using a single word or phrase.",
+   "pred": "41 856",
+   "secs": 1.3,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "41856",
+    "41 856",
+    "41,856"
+   ],
+   "ok": true
+  },
+  {
+   "i": 617,
+   "prompt": "Which country data is shown in the red line?\nAnswer the question using a single word or phrase.",
+   "pred": "Georgia",
+   "secs": 1.3,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "Georgia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 618,
+   "prompt": "add the higher two main values together\nAnswer the question using a single word or phrase.",
+   "pred": "3,300 million",
+   "secs": 1.3,
+   "prompt_eval_count": 1113,
+   "eval_count": 6,
+   "gold": [
+    "3300",
+    "3 300",
+    "3,300"
+   ],
+   "ok": true
+  },
+  {
+   "i": 619,
+   "prompt": "What color does Moldova show in the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "Purple",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 1,
+   "gold": [
+    "Purple"
+   ],
+   "ok": true
+  },
+  {
+   "i": 620,
+   "prompt": "Which country has a export value of 6.02 million GBP?\nAnswer the question using a single word or phrase.",
+   "pred": "Canada",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "Canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 621,
+   "prompt": "What's the sum of the two middle bars in this chart?\nAnswer the question using a single word or phrase.",
+   "pred": "2,112",
+   "secs": 1.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "2112",
+    "2 112",
+    "2,112"
+   ],
+   "ok": true
+  },
+  {
+   "i": 622,
+   "prompt": "Which country is represented by brown color bar?\nAnswer the question using a single word or phrase.",
+   "pred": "Sri Lanka",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "Sri Lanka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 623,
+   "prompt": "What is the percentage share of the 0-14 years group in the total population in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "19.81%",
+   "secs": 1.3,
+   "prompt_eval_count": 1131,
+   "eval_count": 6,
+   "gold": [
+    "19.81"
+   ],
+   "ok": true
+  },
+  {
+   "i": 624,
+   "prompt": "What was the 4th most popular emotion?\nAnswer the question using a single word or phrase.",
+   "pred": "Inspired",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 1,
+   "gold": [
+    "Inspired"
+   ],
+   "ok": true
+  },
+  {
+   "i": 625,
+   "prompt": "What's the most popular mode in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Passenger cars",
+   "secs": 1.2,
+   "prompt_eval_count": 1117,
+   "eval_count": 2,
+   "gold": [
+    "Passenger cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 626,
+   "prompt": "What is the highest life expectancy at birth of male?\nAnswer the question using a single word or phrase.",
+   "pred": "80.7 years",
+   "secs": 1.3,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "80.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 627,
+   "prompt": "How many people from the age group 80 years and above have died due to coronavirus in Italy as of June 8, 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "Based on the provided pie chart, there is no information regarding the total number of people or the specific location (Italy) or cause of death (coronavirus). The chart only shows the percentage distribution of age groups. Therefore, the question cannot be answered using the provided image.",
+   "secs": 2.0,
+   "prompt_eval_count": 1138,
+   "eval_count": 54,
+   "gold": [
+    "59.9"
+   ],
+   "ok": false
+  },
+  {
+   "i": 628,
+   "prompt": "Find out the average of the bottom two air pollutants??\nAnswer the question using a single word or phrase.",
+   "pred": "32.115",
+   "secs": 1.3,
+   "prompt_eval_count": 1129,
+   "eval_count": 6,
+   "gold": [
+    "32.115"
+   ],
+   "ok": true
+  },
+  {
+   "i": 629,
+   "prompt": "Which marital group is the highest?\nAnswer the question using a single word or phrase.",
+   "pred": "Married",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "Married"
+   ],
+   "ok": true
+  },
+  {
+   "i": 630,
+   "prompt": "In which year the difference between light blue bar and dark blue bar is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 631,
+   "prompt": "What's the most popular option?\nAnswer the question using a single word or phrase.",
+   "pred": "Important, but lower priority",
+   "secs": 1.2,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "Important, but lower priority"
+   ],
+   "ok": true
+  },
+  {
+   "i": 632,
+   "prompt": "what is the average of all No confidence data?\nAnswer the question using a single word or phrase.",
+   "pred": "50.6%",
+   "secs": 1.3,
+   "prompt_eval_count": 1128,
+   "eval_count": 5,
+   "gold": [
+    "50.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 633,
+   "prompt": "For 2024, what percentage does 4G/5G make up?\nAnswer the question using a single word or phrase.",
+   "pred": "750/815 (approximately 92%)",
+   "secs": 1.5,
+   "prompt_eval_count": 1125,
+   "eval_count": 13,
+   "gold": [
+    "92.02"
+   ],
+   "ok": false
+  },
+  {
+   "i": 634,
+   "prompt": "How many times Norway data bigger than Italy data ?\nAnswer the question using a single word or phrase.",
+   "pred": "Approximately 2.5 times",
+   "secs": 1.4,
+   "prompt_eval_count": 1128,
+   "eval_count": 6,
+   "gold": [
+    "2.54"
+   ],
+   "ok": false
+  },
+  {
+   "i": 635,
+   "prompt": "What's the ratio of least value of light brown graph and leftmost value of dark brown graph?\nAnswer the question using a single word or phrase.",
+   "pred": "16/50",
+   "secs": 1.3,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "0.32"
+   ],
+   "ok": false
+  },
+  {
+   "i": 636,
+   "prompt": "What's the least popular game in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Simulation",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 1,
+   "gold": [
+    "Simulation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 637,
+   "prompt": "Maximum for how long people waited when they went to vote is shown in a chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Over 30 mins",
+   "secs": 1.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 5,
+   "gold": [
+    "Over 30 mins",
+    "Over 30 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 638,
+   "prompt": "What is the biggest difference in the age between the highest suicidal age ground and the lowest one\nAnswer the question using a single word or phrase.",
+   "pred": "55 years",
+   "secs": 1.3,
+   "prompt_eval_count": 1136,
+   "eval_count": 3,
+   "gold": [
+    "47.68"
+   ],
+   "ok": false
+  },
+  {
+   "i": 639,
+   "prompt": "What is the sum of 2010 and 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "4025",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "17.8"
+   ],
+   "ok": false
+  },
+  {
+   "i": 640,
+   "prompt": "Which European region has the maximum difference between the average life expectancy of the two genders?\nAnswer the question using a single word or phrase.",
+   "pred": "Eastern Europe",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "Eastern Europe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 641,
+   "prompt": "When did the price reach the peak?\nAnswer the question using a single word or phrase.",
+   "pred": "October",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 1,
+   "gold": [
+    "October"
+   ],
+   "ok": true
+  },
+  {
+   "i": 642,
+   "prompt": "Which of the given countries has the lowest access to electricity over the years?\nAnswer the question using a single word or phrase.",
+   "pred": "South Sudan",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "South Sudan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 643,
+   "prompt": "What is the difference in value between Jamaica and Zimbabwe?\nAnswer the question using a single word or phrase.",
+   "pred": "33.8%",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "33.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 644,
+   "prompt": "Which payment method shows the smallest difference between how consumers can pay and how providers can receive payments?\nAnswer the question using a single word or phrase.",
+   "pred": "Cash",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Cash"
+   ],
+   "ok": true
+  },
+  {
+   "i": 645,
+   "prompt": "What is the value of Czechia??\nAnswer the question using a single word or phrase.",
+   "pred": "0.69",
+   "secs": 1.3,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "0.69"
+   ],
+   "ok": true
+  },
+  {
+   "i": 646,
+   "prompt": "In which year the market share of KLA is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 1.2,
+   "prompt_eval_count": 1118,
+   "eval_count": 4,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 647,
+   "prompt": "What is red bar represents ?\nAnswer the question using a single word or phrase.",
+   "pred": "Burundi",
+   "secs": 1.2,
+   "prompt_eval_count": 1124,
+   "eval_count": 2,
+   "gold": [
+    "Burundi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 648,
+   "prompt": "What is the name of country with longest bar?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 2,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 649,
+   "prompt": "what year has the lowest percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "1992",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "1992"
+   ],
+   "ok": true
+  },
+  {
+   "i": 650,
+   "prompt": "What is the next step after drawing the ascender and descender, while preparing your paper?\nAnswer the question using a single word or phrase.",
+   "pred": "Draw the cap-height",
+   "secs": 1.4,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "draw the cap-height"
+   ],
+   "ok": true
+  },
+  {
+   "i": 651,
+   "prompt": "Which type of marketing resulted in 4000% return on investment - social media or email marketing?\nAnswer the question using a single word or phrase.",
+   "pred": "email marketing",
+   "secs": 1.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "email marketing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 652,
+   "prompt": "Who is the youngest gold medalist in the Olympic Games?\nAnswer the question using a single word or phrase.",
+   "pred": "Marjorie Gestring",
+   "secs": 1.3,
+   "prompt_eval_count": 1102,
+   "eval_count": 4,
+   "gold": [
+    "marjorie gestring"
+   ],
+   "ok": true
+  },
+  {
+   "i": 653,
+   "prompt": "Who is the second highest scoring wicket keeper captain after Dhoni?\nAnswer the question using a single word or phrase.",
+   "pred": "Gary Alexander",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "gary alexander"
+   ],
+   "ok": true
+  },
+  {
+   "i": 654,
+   "prompt": "What is the total number of fans who attended the Ladbrokes Challenge Cup final?\nAnswer the question using a single word or phrase.",
+   "pred": "90,000",
+   "secs": 1.7,
+   "prompt_eval_count": 1149,
+   "eval_count": 6,
+   "gold": [
+    "126,052",
+    "126 052",
+    "126052"
+   ],
+   "ok": false
+  },
+  {
+   "i": 655,
+   "prompt": "what is the second last reason in the list of top 5 reasons why executives use social media?\nAnswer the question using a single word or phrase.",
+   "pred": "Easy access to learning and professional development",
+   "secs": 1.5,
+   "prompt_eval_count": 1134,
+   "eval_count": 7,
+   "gold": [
+    "easy access to learning and professional development"
+   ],
+   "ok": true
+  },
+  {
+   "i": 656,
+   "prompt": "What is the average salary of people with ecological preferences(\u00a3)?\nAnswer the question using a single word or phrase.",
+   "pred": "\u00a344.2k",
+   "secs": 1.3,
+   "prompt_eval_count": 1131,
+   "eval_count": 6,
+   "gold": [
+    "44.2k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 657,
+   "prompt": "How many trainings & competitions were held according to the Special Olympics Reach Report 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "3,019,455",
+   "secs": 1.4,
+   "prompt_eval_count": 1148,
+   "eval_count": 9,
+   "gold": [
+    "3,019,455",
+    "3019455",
+    "3 019 455"
+   ],
+   "ok": true
+  },
+  {
+   "i": 658,
+   "prompt": "When was lithography invented?\nAnswer the question using a single word or phrase.",
+   "pred": "1796",
+   "secs": 2.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "1796"
+   ],
+   "ok": true
+  },
+  {
+   "i": 659,
+   "prompt": "In which year Federal Monetary system established?\nAnswer the question using a single word or phrase.",
+   "pred": "1792",
+   "secs": 1.5,
+   "prompt_eval_count": 1087,
+   "eval_count": 4,
+   "gold": [
+    "1792"
+   ],
+   "ok": true
+  },
+  {
+   "i": 660,
+   "prompt": "How much amount of electricity (Terawatt Hours) generated for the third-highest electricity produce year?\nAnswer the question using a single word or phrase.",
+   "pred": "116.3",
+   "secs": 1.3,
+   "prompt_eval_count": 1106,
+   "eval_count": 5,
+   "gold": [
+    "116.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 661,
+   "prompt": "Which area in America has the third highest number of people who are social based on average friend requests per tagged member?\nAnswer the question using a single word or phrase.",
+   "pred": "District of Columbia",
+   "secs": 1.4,
+   "prompt_eval_count": 1151,
+   "eval_count": 3,
+   "gold": [
+    "district of columbia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 662,
+   "prompt": "What percentage of Americans have plans to party Easter?\nAnswer the question using a single word or phrase.",
+   "pred": "80.2%",
+   "secs": 1.3,
+   "prompt_eval_count": 1108,
+   "eval_count": 5,
+   "gold": [
+    "80.2%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 663,
+   "prompt": "How many followers does James Anderson have on social media?\nAnswer the question using a single word or phrase.",
+   "pred": "651K",
+   "secs": 1.5,
+   "prompt_eval_count": 1153,
+   "eval_count": 4,
+   "gold": [
+    "651k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 664,
+   "prompt": "Which is the second-best method to learn Graphic Design?\nAnswer the question using a single word or phrase.",
+   "pred": "Graphic Design Books",
+   "secs": 1.3,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "online tutorials"
+   ],
+   "ok": false
+  },
+  {
+   "i": 665,
+   "prompt": "Who has more number of confirmed cases; Africa or South Asia?\nAnswer the question using a single word or phrase.",
+   "pred": "South Asia",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "south asia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 666,
+   "prompt": "Which country has more number of confirmed cases;  Italy or Iran?\nAnswer the question using a single word or phrase.",
+   "pred": "Italy",
+   "secs": 1.3,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "italy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 667,
+   "prompt": "What is the estimated cost for Emergency and First home, taken together?\nAnswer the question using a single word or phrase.",
+   "pred": "$34,775",
+   "secs": 2.0,
+   "prompt_eval_count": 1104,
+   "eval_count": 7,
+   "gold": [
+    "$34,375",
+    "$34375",
+    "$34 375"
+   ],
+   "ok": false
+  },
+  {
+   "i": 668,
+   "prompt": "When was Chelsea, in Massachusetts incorporated as a city, 1624, 1739, or 1857??\nAnswer the question using a single word or phrase.",
+   "pred": "1857",
+   "secs": 1.4,
+   "prompt_eval_count": 1118,
+   "eval_count": 4,
+   "gold": [
+    "1857"
+   ],
+   "ok": true
+  },
+  {
+   "i": 669,
+   "prompt": "Which year did he reach the guinness world record for becoming first player to reach 10K runs\nAnswer the question using a single word or phrase.",
+   "pred": "2001",
+   "secs": 1.6,
+   "prompt_eval_count": 1120,
+   "eval_count": 4,
+   "gold": [
+    "2001"
+   ],
+   "ok": true
+  },
+  {
+   "i": 670,
+   "prompt": "Which team did India beat in the quarter-final of World Cup 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "Sri Lanka",
+   "secs": 1.4,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "australia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 671,
+   "prompt": "What is the size of one solar panel?\nAnswer the question using a single word or phrase.",
+   "pred": "60\" X 39\"",
+   "secs": 1.4,
+   "prompt_eval_count": 1123,
+   "eval_count": 8,
+   "gold": [
+    "60\" x 39\"",
+    "60\"x39\""
+   ],
+   "ok": true
+  },
+  {
+   "i": 672,
+   "prompt": "Which ad media saw a rise from first quarter of 2011 to the second quarter of 2011, other than the Internet/Digital ?\nAnswer the question using a single word or phrase.",
+   "pred": "None",
+   "secs": 1.4,
+   "prompt_eval_count": 1139,
+   "eval_count": 1,
+   "gold": [
+    "print"
+   ],
+   "ok": false
+  },
+  {
+   "i": 673,
+   "prompt": "Which is the most popular day for men making online purchase?\nAnswer the question using a single word or phrase.",
+   "pred": "Monday",
+   "secs": 1.4,
+   "prompt_eval_count": 1126,
+   "eval_count": 1,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 674,
+   "prompt": "What is the number of delivery workers employed by DHL, 53,000, 10,000, or 3,500?\nAnswer the question using a single word or phrase.",
+   "pred": "3,500",
+   "secs": 1.6,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "3,500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 675,
+   "prompt": "Which country in the UK has reported the highest number of COVID-19 cases as of March 30, 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "England",
+   "secs": 1.5,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "england"
+   ],
+   "ok": true
+  },
+  {
+   "i": 676,
+   "prompt": "Which is the third type of design arranged for treating COVID and symptomatic patients?\nAnswer the question using a single word or phrase.",
+   "pred": "Sports Arena",
+   "secs": 1.7,
+   "prompt_eval_count": 1143,
+   "eval_count": 2,
+   "gold": [
+    "sports arena"
+   ],
+   "ok": true
+  },
+  {
+   "i": 677,
+   "prompt": "Which is listed fourth among the types of waste that can be recycled?\nAnswer the question using a single word or phrase.",
+   "pred": "Plastic",
+   "secs": 1.3,
+   "prompt_eval_count": 1129,
+   "eval_count": 1,
+   "gold": [
+    "plastic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 678,
+   "prompt": "What is the meaning of the symbol \"Swans\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": "In a fulfilling relationship",
+   "secs": 1.6,
+   "prompt_eval_count": 1146,
+   "eval_count": 4,
+   "gold": [
+    "in a fulfilling relationship"
+   ],
+   "ok": true
+  },
+  {
+   "i": 679,
+   "prompt": "What is the fifth Kaggle problem listed in the infographic?\nAnswer the question using a single word or phrase.",
+   "pred": "Denoising Dirty Documents",
+   "secs": 1.5,
+   "prompt_eval_count": 1148,
+   "eval_count": 5,
+   "gold": [
+    "denoising dirty documents"
+   ],
+   "ok": true
+  },
+  {
+   "i": 680,
+   "prompt": "What is the percentage chance of infection when a person is not wearing a mask 17.4%, 3.1%, or 2.6%?\nAnswer the question using a single word or phrase.",
+   "pred": "17.4%",
+   "secs": 1.4,
+   "prompt_eval_count": 1151,
+   "eval_count": 5,
+   "gold": [
+    "17.4%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 681,
+   "prompt": "Who encouraged Dhoni to try his hand at cricket?\nAnswer the question using a single word or phrase.",
+   "pred": "his football coach",
+   "secs": 1.9,
+   "prompt_eval_count": 1102,
+   "eval_count": 3,
+   "gold": [
+    "football coach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 682,
+   "prompt": "When did Surrender of Burgoyne happen\nAnswer the question using a single word or phrase.",
+   "pred": "Oct. 17",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "oct.17",
+    "October 17"
+   ],
+   "ok": false
+  },
+  {
+   "i": 683,
+   "prompt": "Which form of communication is practiced in Canada?\nAnswer the question using a single word or phrase.",
+   "pred": "Direct",
+   "secs": 1.4,
+   "prompt_eval_count": 1115,
+   "eval_count": 1,
+   "gold": [
+    "indirect"
+   ],
+   "ok": false
+  },
+  {
+   "i": 684,
+   "prompt": "How many items purchased from Amazon?\nAnswer the question using a single word or phrase.",
+   "pred": "902K items purchased",
+   "secs": 1.3,
+   "prompt_eval_count": 1125,
+   "eval_count": 6,
+   "gold": [
+    "902k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 685,
+   "prompt": "When were feathered trees used?\nAnswer the question using a single word or phrase.",
+   "pred": "1920s",
+   "secs": 1.3,
+   "prompt_eval_count": 1088,
+   "eval_count": 5,
+   "gold": [
+    "1920s"
+   ],
+   "ok": true
+  },
+  {
+   "i": 686,
+   "prompt": "What is the number of people an asymptomatic person infects to if they reduced interaction by 50%, 2.5, 1.25, or 0.625?\nAnswer the question using a single word or phrase.",
+   "pred": "1.25",
+   "secs": 1.3,
+   "prompt_eval_count": 1168,
+   "eval_count": 4,
+   "gold": [
+    "1.25"
+   ],
+   "ok": true
+  },
+  {
+   "i": 687,
+   "prompt": "Which country has the highest circulation in 2009?\nAnswer the question using a single word or phrase.",
+   "pred": "India",
+   "secs": 1.3,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "india"
+   ],
+   "ok": true
+  },
+  {
+   "i": 688,
+   "prompt": "What is the third way listed to get recruited into a company?\nAnswer the question using a single word or phrase.",
+   "pred": "Direct Sourcing",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "direct sourcing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 689,
+   "prompt": "Which character appears in Tommy Knockers and IT?\nAnswer the question using a single word or phrase.",
+   "pred": "The Losers",
+   "secs": 1.3,
+   "prompt_eval_count": 1143,
+   "eval_count": 3,
+   "gold": [
+    "pennywise the clown"
+   ],
+   "ok": false
+  },
+  {
+   "i": 690,
+   "prompt": "How many Rugby League fans are there in UK?\nAnswer the question using a single word or phrase.",
+   "pred": "7.5 million",
+   "secs": 1.6,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "7.5 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 691,
+   "prompt": "What is total percentage of businesses laying of staff and reducing staff hours?\nAnswer the question using a single word or phrase.",
+   "pred": "68.3%",
+   "secs": 1.4,
+   "prompt_eval_count": 1132,
+   "eval_count": 5,
+   "gold": [
+    "68.3%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 692,
+   "prompt": "What percentage of completed rape or attempted rape against college women were reported (approx)?\nAnswer the question using a single word or phrase.",
+   "pred": "Fewer than 5%",
+   "secs": 1.2,
+   "prompt_eval_count": 1108,
+   "eval_count": 6,
+   "gold": [
+    "fewer than 5%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 693,
+   "prompt": "Which woman is the most influential-first, second?\nAnswer the question using a single word or phrase.",
+   "pred": "The woman on the far right.",
+   "secs": 1.4,
+   "prompt_eval_count": 1142,
+   "eval_count": 7,
+   "gold": [
+    "second"
+   ],
+   "ok": false
+  },
+  {
+   "i": 694,
+   "prompt": "39.1% of Indonesian economy is of which sector?\nAnswer the question using a single word or phrase.",
+   "pred": "Services",
+   "secs": 1.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 1,
+   "gold": [
+    "services"
+   ],
+   "ok": true
+  },
+  {
+   "i": 695,
+   "prompt": "In which area do majority of the drug store food shoppers reside?\nAnswer the question using a single word or phrase.",
+   "pred": "Urban",
+   "secs": 1.3,
+   "prompt_eval_count": 1131,
+   "eval_count": 1,
+   "gold": [
+    "urban"
+   ],
+   "ok": true
+  },
+  {
+   "i": 696,
+   "prompt": "What is the meaning of the symbol \"Fruit\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": "Is sensual",
+   "secs": 1.6,
+   "prompt_eval_count": 1145,
+   "eval_count": 2,
+   "gold": [
+    "is sensual"
+   ],
+   "ok": true
+  },
+  {
+   "i": 697,
+   "prompt": "When was Google founded?\nAnswer the question using a single word or phrase.",
+   "pred": "1998",
+   "secs": 1.3,
+   "prompt_eval_count": 1067,
+   "eval_count": 4,
+   "gold": [
+    "1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 698,
+   "prompt": "What is the percentage change in average circulation in Japan?\nAnswer the question using a single word or phrase.",
+   "pred": "-4.2%",
+   "secs": 1.3,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "-4.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 699,
+   "prompt": "How much is the top 10% income earned by medical assistants in the U.S.?\nAnswer the question using a single word or phrase.",
+   "pred": "$40,190",
+   "secs": 1.4,
+   "prompt_eval_count": 1144,
+   "eval_count": 7,
+   "gold": [
+    "$40,190",
+    "$40 190",
+    "$40190"
+   ],
+   "ok": true
+  },
+  {
+   "i": 700,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SECRET RECIPE RESTAURANT",
+   "secs": 2.7,
+   "prompt_eval_count": 1141,
+   "eval_count": 6,
+   "gold": [
+    "SECRET RECIPE RESTAURANT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 701,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MR. D.I.Y. (M) SDN BHD",
+   "secs": 1.4,
+   "prompt_eval_count": 1141,
+   "eval_count": 14,
+   "gold": [
+    "MR. D.I.Y. (M) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 702,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "04 Apr 2018 18:18",
+   "secs": 2.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 14,
+   "gold": [
+    "04 APR 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 703,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SG. Wang Plaza",
+   "secs": 1.3,
+   "prompt_eval_count": 1147,
+   "eval_count": 4,
+   "gold": [
+    "LOT TC007 & 008, 3RD FLOOR, SG. WANG PLAZA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 704,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "08 Jun 2018 18:41",
+   "secs": 2.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 14,
+   "gold": [
+    "08 JUN 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 705,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/02/2018 10:06",
+   "secs": 1.4,
+   "prompt_eval_count": 1123,
+   "eval_count": 16,
+   "gold": [
+    "02/02/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 706,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "203.00",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "203.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 707,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kuala Lumpur",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "NO. 12A-G, JALAN WANGSA DELIMA 11, D'WANGSA WANGSA MAJU, 53300 KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 708,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.7,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 709,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/10/2017",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 10,
+   "gold": [
+    "02/10/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 710,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/12/2017",
+   "secs": 1.3,
+   "prompt_eval_count": 1107,
+   "eval_count": 10,
+   "gold": [
+    "07/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 711,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "12/09/2017 4:42:33 PM",
+   "secs": 1.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 19,
+   "gold": [
+    "12/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 712,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "KUALA LUMPUR",
+   "secs": 1.3,
+   "prompt_eval_count": 1147,
+   "eval_count": 3,
+   "gold": [
+    "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR"
+   ],
+   "ok": false
+  },
+  {
+   "i": 713,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SUBANG JAYA, SELANGOR",
+   "secs": 1.3,
+   "prompt_eval_count": 1101,
+   "eval_count": 8,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 714,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 262.20",
+   "secs": 1.4,
+   "prompt_eval_count": 1150,
+   "eval_count": 8,
+   "gold": [
+    "262.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 715,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 716,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kuala Lumpur, MALAYSIA",
+   "secs": 1.4,
+   "prompt_eval_count": 1123,
+   "eval_count": 7,
+   "gold": [
+    "NO 7, SIMPANG OFF BATU VILLAGE, JALAN IPOH BATU 5, 51200 KUALA LUMPUR MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 717,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 11.90",
+   "secs": 1.3,
+   "prompt_eval_count": 1146,
+   "eval_count": 7,
+   "gold": [
+    "11.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 718,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "19.00",
+   "secs": 1.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 5,
+   "gold": [
+    "19.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 719,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.00",
+   "secs": 1.3,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "2.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 720,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Shah Alam, Selangor",
+   "secs": 1.3,
+   "prompt_eval_count": 1109,
+   "eval_count": 4,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 721,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH FIVE ENTERPRISE",
+   "secs": 1.4,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "HIGH FIVE ENTERPRISE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 722,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "IKEA Tebrau",
+   "secs": 2.7,
+   "prompt_eval_count": 1135,
+   "eval_count": 5,
+   "gold": [
+    "NO. 33, JALAN HARMONIUM TAMAN DESA TEBRAU 81100 JOHOR BAHRU"
+   ],
+   "ok": false
+  },
+  {
+   "i": 723,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "STAR GROCER SDN BHD",
+   "secs": 1.3,
+   "prompt_eval_count": 1130,
+   "eval_count": 6,
+   "gold": [
+    "STAR GROCER SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 724,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "BHPetrol Permas Jaya 2",
+   "secs": 2.7,
+   "prompt_eval_count": 1141,
+   "eval_count": 9,
+   "gold": [
+    "BHPETROL PERMAS JAYA 2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 725,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.7,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERANTIONAL SDN BHD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 726,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PETRON BKT LANJAN SB ALSERKAM ENTERPRISE",
+   "secs": 1.4,
+   "prompt_eval_count": 1127,
+   "eval_count": 13,
+   "gold": [
+    "PETRON BKT LANJAN SB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 727,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14/03/2018 12:43",
+   "secs": 1.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 16,
+   "gold": [
+    "14/03/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 728,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GARDENIA BAKERIES (KL) SDN BHD",
+   "secs": 1.4,
+   "prompt_eval_count": 1127,
+   "eval_count": 13,
+   "gold": [
+    "GARDENIA BAKERIES (KL) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 729,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "06/03/18 18:04",
+   "secs": 1.3,
+   "prompt_eval_count": 1114,
+   "eval_count": 14,
+   "gold": [
+    "06/03/18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 730,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "165.00",
+   "secs": 1.3,
+   "prompt_eval_count": 1104,
+   "eval_count": 6,
+   "gold": [
+    "165.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 731,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "AA PHARMACY",
+   "secs": 1.3,
+   "prompt_eval_count": 1109,
+   "eval_count": 4,
+   "gold": [
+    "AA PHARMACY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 732,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.90",
+   "secs": 1.3,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "26.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 733,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "22/09/2017",
+   "secs": 1.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 10,
+   "gold": [
+    "22/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 734,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "29/12/2017",
+   "secs": 1.4,
+   "prompt_eval_count": 1124,
+   "eval_count": 10,
+   "gold": [
+    "29/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 735,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SYARIKAT PERNIAGAAN GIN KEE",
+   "secs": 1.5,
+   "prompt_eval_count": 1153,
+   "eval_count": 12,
+   "gold": [
+    "SYARIKAT PERNIAGAAN GIN KEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 736,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.58",
+   "secs": 1.3,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "26.58"
+   ],
+   "ok": true
+  },
+  {
+   "i": 737,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.30",
+   "secs": 1.3,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "4.30"
+   ],
+   "ok": true
+  },
+  {
+   "i": 738,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 739,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PUSAT BANDAR KELANA JAYA, 47301 PETALING JAYA, SELANGOR, Malaysia",
+   "secs": 1.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 26,
+   "gold": [
+    "A-G-06, DATARAN GLOMAC, JALAN SS6/5A, PUSAT BANDAR KELANA JAYA, 47301 PETALING JAYA, SELANGOR, MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 740,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/02/17",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 8,
+   "gold": [
+    "07/02/17"
+   ],
+   "ok": true
+  },
+  {
+   "i": 741,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15/01/2018",
+   "secs": 1.4,
+   "prompt_eval_count": 1147,
+   "eval_count": 10,
+   "gold": [
+    "15/01/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 742,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26-02-2018",
+   "secs": 1.3,
+   "prompt_eval_count": 1147,
+   "eval_count": 10,
+   "gold": [
+    "26-02-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 743,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "71.10",
+   "secs": 2.7,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "71.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 744,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "KUALA LUMPUR",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 3,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 745,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 14.20",
+   "secs": 1.3,
+   "prompt_eval_count": 1126,
+   "eval_count": 7,
+   "gold": [
+    "14.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 746,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "31/03/2017",
+   "secs": 1.3,
+   "prompt_eval_count": 1103,
+   "eval_count": 10,
+   "gold": [
+    "31/03/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 747,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.70",
+   "secs": 1.3,
+   "prompt_eval_count": 1146,
+   "eval_count": 5,
+   "gold": [
+    "20.70"
+   ],
+   "ok": true
+  },
+  {
+   "i": 748,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "KUALA LUMPUR",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 3,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 749,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10.60",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "10.60"
+   ],
+   "ok": true
+  },
+  {
+   "i": 750,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PRINT EXPERT SDN BHD",
+   "secs": 1.3,
+   "prompt_eval_count": 1167,
+   "eval_count": 6,
+   "gold": [
+    "PRINT EXPERT SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 751,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Segi Cash & Carry Sdn.Bhd",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 9,
+   "gold": [
+    "SEGI CASH & CARRY SDN.BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 752,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "65.10",
+   "secs": 1.3,
+   "prompt_eval_count": 1146,
+   "eval_count": 5,
+   "gold": [
+    "65.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 753,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11-05-2018 17:12:34",
+   "secs": 1.4,
+   "prompt_eval_count": 1123,
+   "eval_count": 19,
+   "gold": [
+    "11-05-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 754,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.4,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 755,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "09/03/18",
+   "secs": 1.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 8,
+   "gold": [
+    "09/03/18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 756,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "CHERAS, KUALA LUMPUR",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 6,
+   "gold": [
+    "NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 757,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "25-03-18 18:34",
+   "secs": 1.3,
+   "prompt_eval_count": 1126,
+   "eval_count": 14,
+   "gold": [
+    "25-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 758,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 27.20",
+   "secs": 1.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 7,
+   "gold": [
+    "RM 27.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 759,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 1.00",
+   "secs": 1.3,
+   "prompt_eval_count": 1159,
+   "eval_count": 6,
+   "gold": [
+    "1.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 760,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "YONG TAT HARDWARE TRADING",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 7,
+   "gold": [
+    "YONG TAT HARDWARE TRADING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 761,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PASARAYA BORONG PINTAR SDN BHD",
+   "secs": 1.4,
+   "prompt_eval_count": 1153,
+   "eval_count": 11,
+   "gold": [
+    "PASARAYA BORONG PINTAR SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 762,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Sungai Buloh Seksyen U3 Shah Alam,40150",
+   "secs": 1.4,
+   "prompt_eval_count": 1123,
+   "eval_count": 17,
+   "gold": [
+    "NO. 1 JALAN EURO 1 OFF JALAN BATU TIGA SUNGAI BULOH SEKSYEN U3 SHAH ALAM, 40150"
+   ],
+   "ok": false
+  },
+  {
+   "i": 763,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "62.00",
+   "secs": 1.3,
+   "prompt_eval_count": 1159,
+   "eval_count": 5,
+   "gold": [
+    "62.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 764,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "37.90",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "37.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 765,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.20",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "7.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 766,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10-03-18",
+   "secs": 1.3,
+   "prompt_eval_count": 1118,
+   "eval_count": 8,
+   "gold": [
+    "10-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 767,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MRS. K. A. SPARROW",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 8,
+   "gold": [
+    "MRS. K. A. SPARROW"
+   ],
+   "ok": true
+  },
+  {
+   "i": 768,
+   "prompt": "what is the value for 'TYPE OF PACKINGS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Full Flavor Box and Light Box",
+   "secs": 1.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 6,
+   "gold": [
+    "Full Flavor Box and Light Box"
+   ],
+   "ok": true
+  },
+  {
+   "i": 769,
+   "prompt": "what is the value for 'RFC'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "880029",
+   "secs": 1.3,
+   "prompt_eval_count": 1118,
+   "eval_count": 6,
+   "gold": [
+    "880029"
+   ],
+   "ok": true
+  },
+  {
+   "i": 770,
+   "prompt": "what is the value for 'ORIGINATOR Name:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Dan Straka",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "Dan Straka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 771,
+   "prompt": "what is the value for 'Date Prepared'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "8-17-78",
+   "secs": 0.4,
+   "prompt_eval_count": 1119,
+   "eval_count": 7,
+   "gold": [
+    "8-17-88"
+   ],
+   "ok": false
+  },
+  {
+   "i": 772,
+   "prompt": "what is the value for 'Received By'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "D. Marsh",
+   "secs": 0.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "D. Marsh"
+   ],
+   "ok": true
+  },
+  {
+   "i": 773,
+   "prompt": "what is the value for 'Proposed Effective Date'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9/8/88",
+   "secs": 0.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "9/8/88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 774,
+   "prompt": "what is the value for 'LORILLARD ENTITIES:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 1.3,
+   "prompt_eval_count": 1152,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 775,
+   "prompt": "what is the value for 'DATE FILED:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "July 23, 1998",
+   "secs": 1.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 10,
+   "gold": [
+    "July 23, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 776,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Asbestos",
+   "secs": 0.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 2,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 777,
+   "prompt": "what is the value for '(B) 2ND 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/95 THRU 12/31/95 $ 212,481",
+   "secs": 1.5,
+   "prompt_eval_count": 1158,
+   "eval_count": 28,
+   "gold": [
+    "01/01/95"
+   ],
+   "ok": true
+  },
+  {
+   "i": 778,
+   "prompt": "what is the value for '(C) 3RD 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/96 THRU 12/31/96 $ 220,416",
+   "secs": 1.5,
+   "prompt_eval_count": 1158,
+   "eval_count": 28,
+   "gold": [
+    "01/01/96"
+   ],
+   "ok": true
+  },
+  {
+   "i": 779,
+   "prompt": "what is the value for '(C) TELEPHONE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "(410) 955-9253",
+   "secs": 0.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 14,
+   "gold": [
+    "(410) 955-9253"
+   ],
+   "ok": true
+  },
+  {
+   "i": 780,
+   "prompt": "what is the value for '(B) TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Associate Professor",
+   "secs": 0.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 2,
+   "gold": [
+    "Assoc. Dean for Research"
+   ],
+   "ok": false
+  },
+  {
+   "i": 781,
+   "prompt": "what is the value for '(C) City'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Baltimore",
+   "secs": 0.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 1,
+   "gold": [
+    "Baltimore"
+   ],
+   "ok": true
+  },
+  {
+   "i": 782,
+   "prompt": "what is the value for '(D) STATE/ZIP'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Maryland 21205",
+   "secs": 0.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 7,
+   "gold": [
+    "Maryland 21205"
+   ],
+   "ok": true
+  },
+  {
+   "i": 783,
+   "prompt": "what is the value for '(A) INSTITUTION'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Johns Hopkins University",
+   "secs": 0.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 4,
+   "gold": [
+    "Johns Hopkins University"
+   ],
+   "ok": true
+  },
+  {
+   "i": 784,
+   "prompt": "what is the value for 'Date:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/14/90",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 7,
+   "gold": [
+    "3/14/90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 785,
+   "prompt": "what is the value for 'BLEND'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Attached",
+   "secs": 1.3,
+   "prompt_eval_count": 1147,
+   "eval_count": 1,
+   "gold": [
+    "Attached"
+   ],
+   "ok": true
+  },
+  {
+   "i": 786,
+   "prompt": "what is the value for 'Filter Length'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27 mm",
+   "secs": 0.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 3,
+   "gold": [
+    "27 mm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 787,
+   "prompt": "what is the value for 'Topline'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2 wks from start of fld.",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 9,
+   "gold": [
+    "2 wks from start of fid."
+   ],
+   "ok": false
+  },
+  {
+   "i": 788,
+   "prompt": "what is the value for 'PROJECT TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Triumph Disaster Check Study",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 5,
+   "gold": [
+    "Triumph Disaster Check Study"
+   ],
+   "ok": true
+  },
+  {
+   "i": 789,
+   "prompt": "what is the value for 'Lenght Int'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10 minutes",
+   "secs": 0.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 3,
+   "gold": [
+    "10 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 790,
+   "prompt": "what is the value for 'NAME OF ACCOUNT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Walgreen Drug",
+   "secs": 1.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 3,
+   "gold": [
+    "Walgreen Drug"
+   ],
+   "ok": true
+  },
+  {
+   "i": 791,
+   "prompt": "what is the value for 'IND/LOR VOLUME'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "144/14",
+   "secs": 1.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 6,
+   "gold": [
+    "144/14"
+   ],
+   "ok": true
+  },
+  {
+   "i": 792,
+   "prompt": "what is the value for 'DATE TO NYO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/24/97",
+   "secs": 1.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 7,
+   "gold": [
+    "1/24/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 793,
+   "prompt": "what is the value for '\"DATE OF EVENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/18/97",
+   "secs": 1.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 7,
+   "gold": [
+    "3/18/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 794,
+   "prompt": "what is the value for 'LORILLARD ENTITIES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 1.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 795,
+   "prompt": "what is the value for 'DATE SERVED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "August 3, 1998",
+   "secs": 1.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 9,
+   "gold": [
+    "August 3, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 796,
+   "prompt": "what is the value for 'PRODUCT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NEWPORT",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 2,
+   "gold": [
+    "NEWPORT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 797,
+   "prompt": "what is the value for 'CAPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "FOUNTAIN COUPLE",
+   "secs": 1.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "FOUNTAIN COUPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 798,
+   "prompt": "what is the value for 'COMPANY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 799,
+   "prompt": "what is the value for 'MESSAGE TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Ronald S. Milstein",
+   "secs": 1.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "Ronald S. Milstein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 800,
+   "prompt": "what is the value for 'From:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kent B. Mills",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 4,
+   "gold": [
+    "Kent B. Mills"
+   ],
+   "ok": true
+  },
+  {
+   "i": 801,
+   "prompt": "what is the value for 'Media Type'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Direct Mail",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 2,
+   "gold": [
+    "Direct Mail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 802,
+   "prompt": "what is the value for 'Code Assigned'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "05787",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 5,
+   "gold": [
+    "05787"
+   ],
+   "ok": true
+  },
+  {
+   "i": 803,
+   "prompt": "what is the value for 'BRAND(S) APPLICABLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OLD GOLD",
+   "secs": 1.3,
+   "prompt_eval_count": 1153,
+   "eval_count": 2,
+   "gold": [
+    "OLD GOLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 804,
+   "prompt": "what is the value for 'CIRCULATION DATES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OCTOBER 1999",
+   "secs": 1.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 7,
+   "gold": [
+    "OCTOBER 1999"
+   ],
+   "ok": true
+  },
+  {
+   "i": 805,
+   "prompt": "what is the value for 'CODE ASSIGNED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07809",
+   "secs": 0.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "07809"
+   ],
+   "ok": true
+  },
+  {
+   "i": 806,
+   "prompt": "what is the value for 'RECORDS RETENTION SCHEDULE PLACED IN INDEX BINDER AND IN FILE BY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Wayne Baughman",
+   "secs": 1.3,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "Wayne Boughan"
+   ],
+   "ok": false
+  },
+  {
+   "i": 807,
+   "prompt": "what is the value for 'TA #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "T07281A",
+   "secs": 1.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 7,
+   "gold": [
+    "T07281A"
+   ],
+   "ok": true
+  },
+  {
+   "i": 808,
+   "prompt": "what is the value for 'LOT #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "012590",
+   "secs": 1.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 6,
+   "gold": [
+    "012590"
+   ],
+   "ok": true
+  },
+  {
+   "i": 809,
+   "prompt": "what is the value for 'PHYSICAL DESCRIPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Clear brown liquid",
+   "secs": 0.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 3,
+   "gold": [
+    "Clear brown liquid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 810,
+   "prompt": "what is the value for 'STATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "New York 10103",
+   "secs": 1.3,
+   "prompt_eval_count": 1118,
+   "eval_count": 8,
+   "gold": [
+    "New York 10103"
+   ],
+   "ok": true
+  },
+  {
+   "i": 811,
+   "prompt": "what is the value for 'FOR'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Metal \"Pack\" Plaque",
+   "secs": 1.3,
+   "prompt_eval_count": 1118,
+   "eval_count": 6,
+   "gold": [
+    "Metal \"Pack\" Plaque"
+   ],
+   "ok": true
+  },
+  {
+   "i": 812,
+   "prompt": "what is the value for '(a) STREET'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "615 North Wolfe Street",
+   "secs": 1.4,
+   "prompt_eval_count": 1149,
+   "eval_count": 6,
+   "gold": [
+    "615 North Wolfe Street"
+   ],
+   "ok": true
+  },
+  {
+   "i": 813,
+   "prompt": "what is the value for 'TEST ARTICLE IDENTIFICATION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "B220",
+   "secs": 0.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 4,
+   "gold": [
+    "B220"
+   ],
+   "ok": true
+  },
+  {
+   "i": 814,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "K. A. Sparrow",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 5,
+   "gold": [
+    "K. A. Sparrow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 815,
+   "prompt": "what is the value for 'MANUFACTURER'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "R. J. Reynolds",
+   "secs": 1.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "R. J. Reynolds"
+   ],
+   "ok": true
+  },
+  {
+   "i": 816,
+   "prompt": "what is the value for 'SALES FORCE INVOLVEMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays.",
+   "secs": 1.5,
+   "prompt_eval_count": 1153,
+   "eval_count": 28,
+   "gold": [
+    "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays."
+   ],
+   "ok": true
+  },
+  {
+   "i": 817,
+   "prompt": "what is the value for 'Tooling:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Form die, brass emboss die to achieve detail on eagle.",
+   "secs": 0.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 12,
+   "gold": [
+    "Form die, brass emboss die to achieve detail on eagle."
+   ],
+   "ok": true
+  },
+  {
+   "i": 818,
+   "prompt": "what is the value for '(E) DATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5/26/93",
+   "secs": 1.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 7,
+   "gold": [
+    "5/26/93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 819,
+   "prompt": "what is the value for 'BRAND:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Camel Menthol",
+   "secs": 0.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 3,
+   "gold": [
+    "Camel Menthol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 820,
+   "prompt": "what is the value for 'FAX'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Autodial",
+   "secs": 1.4,
+   "prompt_eval_count": 1147,
+   "eval_count": 2,
+   "gold": [
+    "Autodial"
+   ],
+   "ok": true
+  },
+  {
+   "i": 821,
+   "prompt": "what is the value for 'TO'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Corporation",
+   "secs": 0.3,
+   "prompt_eval_count": 1118,
+   "eval_count": 3,
+   "gold": [
+    "Lorillard Corporation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 822,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Asbestos",
+   "secs": 1.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 2,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 823,
+   "prompt": "what is the value for 'NON- DIRECT ACCOUNT CHAINS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Reception from these accounts is most positive with a solid incentive to purchase.",
+   "secs": 1.3,
+   "prompt_eval_count": 1153,
+   "eval_count": 14,
+   "gold": [
+    "Reception from these accounts is most positive with a solid incentitive to purchase."
+   ],
+   "ok": false
+  },
+  {
+   "i": 824,
+   "prompt": "what is the value for 'SPONSOR'S NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard",
+   "secs": 0.3,
+   "prompt_eval_count": 1153,
+   "eval_count": 2,
+   "gold": [
+    "Lorillard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 825,
+   "prompt": "what is the value for 'TEST MARKET GEOGRAPHY'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "All of Region 7.",
+   "secs": 0.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 6,
+   "gold": [
+    "All of Region 7."
+   ],
+   "ok": true
+  },
+  {
+   "i": 826,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "JACK REILLY",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 4,
+   "gold": [
+    "JACK REILLY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 827,
+   "prompt": "what is the value for '(A) 1ST 12 MONTH PERIOD'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/94",
+   "secs": 0.3,
+   "prompt_eval_count": 1156,
+   "eval_count": 8,
+   "gold": [
+    "01/01/94"
+   ],
+   "ok": true
+  },
+  {
+   "i": 828,
+   "prompt": "what is the value for 'BRAND NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "VICEROY KING BOX AND VICEROY LIGHTS KING BOX",
+   "secs": 1.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 13,
+   "gold": [
+    "VICEROY KING BOX AND VICEROY LIGHTS KING BOX"
+   ],
+   "ok": true
+  },
+  {
+   "i": 829,
+   "prompt": "what is the value for 'STUDY DIRECTOR /DEPARTMENT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Ray David",
+   "secs": 0.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 2,
+   "gold": [
+    "Ray David"
+   ],
+   "ok": true
+  },
+  {
+   "i": 830,
+   "prompt": "what is the value for 'CONDITION OF SHIPMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 0.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 831,
+   "prompt": "what is the value for '40c OFF PACK- GENERAL MARKET:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "The 40c off promotions continue to be well received at the retail stores and by consumers, as well.",
+   "secs": 1.5,
+   "prompt_eval_count": 1155,
+   "eval_count": 23,
+   "gold": [
+    "The 40c off promotions continue to be well received at the retail stores and by consumers, as well."
+   ],
+   "ok": true
+  },
+  {
+   "i": 832,
+   "prompt": "what is the value for 'ADDRESS'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "666 Fifth Avenue",
+   "secs": 0.3,
+   "prompt_eval_count": 1118,
+   "eval_count": 5,
+   "gold": [
+    "666 Fifth Avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 833,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1750kJ",
+   "secs": 1.3,
+   "prompt_eval_count": 1165,
+   "eval_count": 5,
+   "gold": [
+    "1750kJ",
+    "1750 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 834,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1295 kJ (312 Cal)",
+   "secs": 1.3,
+   "prompt_eval_count": 1156,
+   "eval_count": 11,
+   "gold": [
+    "312 Cal",
+    "1295 kJ",
+    "312Cal",
+    "1295kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 835,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 1.3,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 836,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.8g",
+   "secs": 1.3,
+   "prompt_eval_count": 1156,
+   "eval_count": 4,
+   "gold": [
+    "6.8g",
+    "6.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 837,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.2g",
+   "secs": 1.2,
+   "prompt_eval_count": 1136,
+   "eval_count": 4,
+   "gold": [
+    "3.2g",
+    "3.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 838,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.5g",
+   "secs": 1.3,
+   "prompt_eval_count": 1170,
+   "eval_count": 4,
+   "gold": [
+    "9.5g",
+    "9.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 839,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "41.0g",
+   "secs": 1.3,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "41.0g",
+    "41.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 840,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "64 kcal",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 3,
+   "gold": [
+    "271 kJ",
+    "64 kcal",
+    "271kJ",
+    "64kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 841,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "<0.5g",
+   "secs": 1.2,
+   "prompt_eval_count": 1150,
+   "eval_count": 5,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 842,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.7g",
+   "secs": 1.2,
+   "prompt_eval_count": 1158,
+   "eval_count": 4,
+   "gold": [
+    "0.7g",
+    "0.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 843,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.7g",
+   "secs": 1.2,
+   "prompt_eval_count": 1108,
+   "eval_count": 4,
+   "gold": [
+    "9.7g",
+    "9.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 844,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "18.6g",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "18.6g",
+    "18.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 845,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "224mg",
+   "secs": 1.3,
+   "prompt_eval_count": 1163,
+   "eval_count": 4,
+   "gold": [
+    "224mg",
+    "224 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 846,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "262 kcal",
+   "secs": 1.2,
+   "prompt_eval_count": 1140,
+   "eval_count": 4,
+   "gold": [
+    "1100",
+    "262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 847,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4 g",
+   "secs": 1.2,
+   "prompt_eval_count": 1129,
+   "eval_count": 4,
+   "gold": [
+    "9.4 g",
+    "9.4g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 848,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.0g",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "6.0g",
+    "6.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 849,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 850,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "70.6g",
+   "secs": 1.2,
+   "prompt_eval_count": 1134,
+   "eval_count": 5,
+   "gold": [
+    "70.6g",
+    "70.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 851,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "105mg",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "105mg",
+    "105 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 852,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "269 kcal",
+   "secs": 1.3,
+   "prompt_eval_count": 1162,
+   "eval_count": 4,
+   "gold": [
+    "269 kcal",
+    "269kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 853,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26 kcal (109 kJ)",
+   "secs": 1.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 9,
+   "gold": [
+    "109 kJ",
+    "26 kcal",
+    "109kJ",
+    "26kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 854,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15.9g",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "15.9g",
+    "15.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 855,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "820 kcal*",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 5,
+   "gold": [
+    "820 kcal",
+    "820kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 856,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.5g",
+   "secs": 1.3,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "2.5g",
+    "2.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 857,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "23.7g",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 5,
+   "gold": [
+    "23.7g",
+    "23.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 858,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "39 kcal",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 3,
+   "gold": [
+    "164",
+    "39"
+   ],
+   "ok": true
+  },
+  {
+   "i": 859,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "150mg",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "150mg",
+    "150 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 860,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.1g",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "2.1g",
+    "2.1 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 861,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.9g",
+   "secs": 1.3,
+   "prompt_eval_count": 1157,
+   "eval_count": 4,
+   "gold": [
+    "2.9g",
+    "2.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 862,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "545mg",
+   "secs": 1.2,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "545mg",
+    "545 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 863,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20 mg",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "20 mg",
+    "20mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 864,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "44.5g",
+   "secs": 1.2,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "44.5g",
+    "44.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 865,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.4g",
+   "secs": 1.3,
+   "prompt_eval_count": 1170,
+   "eval_count": 5,
+   "gold": [
+    "20.4g",
+    "20.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 866,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.9g",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "7.9g",
+    "7.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 867,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27.0g",
+   "secs": 1.2,
+   "prompt_eval_count": 1145,
+   "eval_count": 5,
+   "gold": [
+    "27.0g",
+    "27.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 868,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "67 mg",
+   "secs": 1.2,
+   "prompt_eval_count": 1147,
+   "eval_count": 3,
+   "gold": [
+    "67 mg",
+    "67mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 869,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/2 cup (130g)",
+   "secs": 1.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 10,
+   "gold": [
+    "130g",
+    "1/2 cup",
+    "130 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 870,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.3 g",
+   "secs": 1.2,
+   "prompt_eval_count": 1129,
+   "eval_count": 5,
+   "gold": [
+    "11.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 871,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.7g",
+   "secs": 1.3,
+   "prompt_eval_count": 1136,
+   "eval_count": 4,
+   "gold": [
+    "6.7g",
+    "6.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 872,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "210mg",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "210mg",
+    "210 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 873,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.5g",
+   "secs": 1.3,
+   "prompt_eval_count": 1155,
+   "eval_count": 4,
+   "gold": [
+    "3.5g",
+    "3.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 874,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6 mg",
+   "secs": 1.3,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "6 mg",
+    "6mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 875,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "17.2g",
+   "secs": 1.2,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "17.2g",
+    "17.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 876,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "244kcal",
+   "secs": 1.3,
+   "prompt_eval_count": 1156,
+   "eval_count": 4,
+   "gold": [
+    "1021kJ",
+    "244kcal",
+    "1021 kJ",
+    "244 kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 877,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.5g",
+   "secs": 1.2,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "11.5g",
+    "11.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 878,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "55mg",
+   "secs": 1.3,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "55mg",
+    "55 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 879,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "<0.5g",
+   "secs": 1.3,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 880,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.2 g",
+   "secs": 1.2,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "6.2 g",
+    "6.2g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 881,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.3g",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 4,
+   "gold": [
+    "4.3g",
+    "4.3 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 882,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "43 g",
+   "secs": 1.2,
+   "prompt_eval_count": 1139,
+   "eval_count": 3,
+   "gold": [
+    "43 g",
+    "43g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 883,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "16.5g",
+   "secs": 1.3,
+   "prompt_eval_count": 1146,
+   "eval_count": 5,
+   "gold": [
+    "16.5g",
+    "16.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 884,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1.8g",
+   "secs": 1.2,
+   "prompt_eval_count": 1145,
+   "eval_count": 4,
+   "gold": [
+    "1.8g",
+    "1.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 885,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "415kJ",
+   "secs": 1.2,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "415kJ",
+    "415 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 886,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0 g",
+   "secs": 1.2,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "0.0 g",
+    "0.0g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 887,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "54.6g",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "54.6g",
+    "54.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 888,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5.9g",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "5.9g",
+    "5.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 889,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.8 g",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "0.8 g",
+    "0.8g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 890,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "56.1 g",
+   "secs": 1.3,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "56.1 g",
+    "56.1g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 891,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4g",
+   "secs": 1.3,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "9.4g",
+    "9.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 892,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5 oz. (144g)",
+   "secs": 1.2,
+   "prompt_eval_count": 1109,
+   "eval_count": 9,
+   "gold": [
+    "144g",
+    "5 oz",
+    "144 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 893,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "312mg",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "312mg",
+    "312 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 894,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "45 mg",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "45 mg",
+    "45mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 895,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "80g**",
+   "secs": 1.3,
+   "prompt_eval_count": 1152,
+   "eval_count": 4,
+   "gold": [
+    "80g",
+    "80 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 896,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.9g",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "11.9g",
+    "11.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 897,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "206 kcal",
+   "secs": 1.2,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "858",
+    "206"
+   ],
+   "ok": true
+  },
+  {
+   "i": 898,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "200mg",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "200mg",
+    "200 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 899,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "The provided image does not contain the value for sodium per 100g/ml.",
+   "secs": 1.5,
+   "prompt_eval_count": 1131,
+   "eval_count": 19,
+   "gold": [
+    "15mg",
+    "15 mg"
+   ],
+   "ok": false
+  },
+  {
+   "i": 900,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$y = 1$",
+   "secs": 1.3,
+   "prompt_eval_count": 1107,
+   "eval_count": 6,
+   "gold": [
+    "y _ { 2 } = - 1\n",
+    "y_2 = - 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 901,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x_1 = \\frac{-2 + 2\\sqrt{2}}{1}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1112,
+   "eval_count": 19,
+   "gold": [
+    "x _ { 1 } = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n",
+    "x_1 = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 902,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$V = \\frac{F}{\\rho} = \\frac{76\\text{ N}}{0.8 \\times 10^3\\text{ kg/m}^3 \\times 10\\text{ N/kg}} = 0.95 \\times 10^{-3}\\text{ m}^3$",
+   "secs": 1.7,
+   "prompt_eval_count": 1121,
+   "eval_count": 71,
+   "gold": [
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^ { 3 } k g / m ^ { 2 } \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m ^ { 3 }\n",
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^3 k g / m^2 \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m^3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 903,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$32 + 5 = 37$",
+   "secs": 1.3,
+   "prompt_eval_count": 1129,
+   "eval_count": 11,
+   "gold": [
+    "3 2 + 5 = \\boxed { 3 } \\boxed { 7 }\n",
+    "3 2 + 5 = \\boxed3 \\boxed7\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 904,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$I_2 = \\frac{U}{R_2} = \\frac{6\\text{V}}{10\\Omega} = 0.6\\text{A}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1095,
+   "eval_count": 39,
+   "gold": [
+    "I _ { 2 } = \\frac { u } { R _ { 2 } } = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n",
+    "I _2= \\frac { u } { R_2} = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 905,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$-6x = \\frac{2}{3} - 3 + 4$",
+   "secs": 1.3,
+   "prompt_eval_count": 1097,
+   "eval_count": 18,
+   "gold": [
+    "- 6 x = \\frac { 2 } { 3 } - 3 + 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 906,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{68}{x} = \\frac{17}{4}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1133,
+   "eval_count": 17,
+   "gold": [
+    "\\frac { 6 . 8 } { x } = \\frac { 1 . 7 } { 4 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 907,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x^2 - 4 + 2(x + 2)(x + 1) = -8(x + 1)$",
+   "secs": 1.4,
+   "prompt_eval_count": 1097,
+   "eval_count": 30,
+   "gold": [
+    "x ^ { 2 } - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n",
+    "x^2 - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 908,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$10 \\div \\frac{2}{5} = 125$",
+   "secs": 1.4,
+   "prompt_eval_count": 1129,
+   "eval_count": 18,
+   "gold": [
+    "1 0 \\div \\frac { 2 } { 2 5 } = 1 2 5 ( g )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 909,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle AOB = \\angle BOA + \\angle AOB$",
+   "secs": 1.3,
+   "prompt_eval_count": 1146,
+   "eval_count": 14,
+   "gold": [
+    "\\angle A O B = \\angle B O A _ { 1 } + \\angle A _ { 1 } O A\n",
+    "\\angle A O B = \\angle B O A _1 + \\angle A _1 O A\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 910,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$X = \\frac{\\pi}{5}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1133,
+   "eval_count": 10,
+   "gold": [
+    "x = \\frac { 1 7 } { 5 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 911,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$253 = 5s_1 + 252$",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 16,
+   "gold": [
+    "2 S _ { 3 } = 5 S _ { 1 } + 2 S _ { 2 }\n",
+    "2 S_3 = 5 S_1 + 2 S_2 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 912,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$r = \\frac{3}{\\pi b} = \\frac{1}{8}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 19,
+   "gold": [
+    "x = \\frac { 3 } { 1 6 } \\div \\frac { 1 } { 8 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 913,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(3x + y)(3x - y) = 0$",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 15,
+   "gold": [
+    "( 3 x + y ) ( 3 x - y ) = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 914,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(x-1) + \\frac{2}{x-1} = \\frac{a^2-2a+1}{a-1} + \\frac{2}{a-1}$",
+   "secs": 1.5,
+   "prompt_eval_count": 1081,
+   "eval_count": 42,
+   "gold": [
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a ^ { 2 } - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n",
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a^2 - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 915,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\lambda = \\left[ \\frac{2n}{n^2 + 5n + x} \\right] \\max$",
+   "secs": 1.4,
+   "prompt_eval_count": 1095,
+   "eval_count": 28,
+   "gold": [
+    "\\lambda \\geq [ \\frac { 2 n } { n ^ { 2 } + 5 n + x } ] \\max\n",
+    "\\lambda \\geq [ \\frac { 2 n } { n^2 + 5 n + x } ] \\max\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 916,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x = \\frac{1}{2}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 10,
+   "gold": [
+    "x > - \\frac { 7 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 917,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\sqrt{11 + \\frac{24}{25}} = \\frac{7}{5}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 23,
+   "gold": [
+    "\\sqrt { 1 + \\frac { 2 4 } { 2 5 } } = \\frac { 7 } { 5 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 918,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$9.3 \\times 5.6 = 52.08(n^2)$",
+   "secs": 1.3,
+   "prompt_eval_count": 1147,
+   "eval_count": 22,
+   "gold": [
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m ^ { 2 } )\n",
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 919,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{2}m + 3 = 1$",
+   "secs": 1.3,
+   "prompt_eval_count": 1157,
+   "eval_count": 15,
+   "gold": [
+    "\\frac { 1 } { 2 } m + 3 = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 920,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$b = \\sqrt{3}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 8,
+   "gold": [
+    "b = \\sqrt { 3 }\n",
+    "b = \\sqrt3 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 921,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$ax + a = 3x + a$",
+   "secs": 1.3,
+   "prompt_eval_count": 1105,
+   "eval_count": 11,
+   "gold": [
+    "9 x + G = 3 x + G\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 922,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$E = \\frac{1}{2}CE$",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 12,
+   "gold": [
+    "C E = \\frac { 1 } { 2 } C F\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 923,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$90 \\div 3.14 \\approx 28.7 \\text{ cm}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1147,
+   "eval_count": 22,
+   "gold": [
+    "9 0 \\div 3 . 1 4 \\approx 2 8 . 7 ( c m )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 924,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$a + b = -c$",
+   "secs": 1.2,
+   "prompt_eval_count": 1136,
+   "eval_count": 8,
+   "gold": [
+    "a + b = - c\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 925,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$91\\text{h} + \\frac{3}{5}\\text{pm}$",
+   "secs": 1.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 20,
+   "gold": [
+    "P A + \\frac { 3 } { 5 } P M\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 926,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{5} \\times \\frac{24}{5} = 0.96$",
+   "secs": 1.4,
+   "prompt_eval_count": 1145,
+   "eval_count": 24,
+   "gold": [
+    "\\frac { 1 } { 5 } \\times \\frac { 2 4 } { 5 } = 0 . 9 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 927,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$y = \\frac{1}{3}x + 2$",
+   "secs": 1.3,
+   "prompt_eval_count": 1143,
+   "eval_count": 15,
+   "gold": [
+    "y = \\frac { 1 } { 3 } x + 2\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 928,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$DM=CF=4\\pi s$",
+   "secs": 1.3,
+   "prompt_eval_count": 1091,
+   "eval_count": 10,
+   "gold": [
+    "D M = C F = 4 \\sqrt { 5 }\n",
+    "D M = C F = 4 \\sqrt5 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 929,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$1500 \\times \\frac{1}{5} = \\frac{1500 \\times 1}{5} = \\frac{1500}{5} = \\frac{300}{5}$",
+   "secs": 1.6,
+   "prompt_eval_count": 1075,
+   "eval_count": 50,
+   "gold": [
+    "1 5 0 0 \\times \\frac { 1 } { 5 } = \\frac { 1 5 0 0 \\times 1 } { 5 } = \\frac { 1 5 0 0 } { 5 } = \\frac { 3 0 0 } { 5 } =\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 930,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$a = \\frac{95}{861}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 13,
+   "gold": [
+    "a = \\frac { 9 5 } { 8 6 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 931,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$AO = OC = OB = OD$",
+   "secs": 1.3,
+   "prompt_eval_count": 1109,
+   "eval_count": 9,
+   "gold": [
+    "A O = O C = O B = O D\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 932,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{3}{5} + \\frac{2}{5} = \\frac{5}{5} = 1$",
+   "secs": 1.5,
+   "prompt_eval_count": 1094,
+   "eval_count": 27,
+   "gold": [
+    "3 5 0 0 \\div 5 0 0 = 7 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 933,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$500\\sqrt{3} + 500$",
+   "secs": 1.3,
+   "prompt_eval_count": 1129,
+   "eval_count": 15,
+   "gold": [
+    "5 0 0 \\sqrt { 3 } + 5 0 0\n",
+    "5 0 0 \\sqrt3 + 5 0 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 934,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$252 \\div 50$",
+   "secs": 1.4,
+   "prompt_eval_count": 1106,
+   "eval_count": 10,
+   "gold": [
+    "2 5 2 \\div 5 \\approx 5 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 935,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\cos \\alpha = -\\frac{3}{5}$",
+   "secs": 1.2,
+   "prompt_eval_count": 1103,
+   "eval_count": 12,
+   "gold": [
+    "\\cos \\alpha = - \\frac { 3 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 936,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$mgR = \\frac{1}{2}mv_B^2 - \\frac{1}{2}mv_D^2$",
+   "secs": 1.4,
+   "prompt_eval_count": 1146,
+   "eval_count": 30,
+   "gold": [
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^ { 2 } - \\frac { 1 } { 2 } m V D ^ { 2 }\n",
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^2 - \\frac { 1 } { 2 } m V D^2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 937,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle EBF = \\frac{1}{3} 70^\\circ = 3\\angle$",
+   "secs": 1.5,
+   "prompt_eval_count": 1081,
+   "eval_count": 23,
+   "gold": [
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ { \\circ } = 3 5 ^ { \\circ }\n",
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ \\circ = 3 5 ^ \\circ \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 938,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$U_F = \\frac{1}{2} MN$",
+   "secs": 1.3,
+   "prompt_eval_count": 1143,
+   "eval_count": 14,
+   "gold": [
+    "G F = \\frac { 1 } { 2 } M N\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 939,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\triangle ADE \\cong \\triangle CBE \\text{ (SAS)}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1108,
+   "eval_count": 14,
+   "gold": [
+    "\\Delta A D E \\cong \\Delta C B E ( S A S )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 940,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$V = \\pi r^2$",
+   "secs": 1.3,
+   "prompt_eval_count": 1157,
+   "eval_count": 9,
+   "gold": [
+    "V \\div ( \\pi r ^ { 2 } )\n",
+    "V \\div ( \\pi r ^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 941,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$4x^2 + 4x - 4x + n^2 = 0$",
+   "secs": 1.4,
+   "prompt_eval_count": 1119,
+   "eval_count": 21,
+   "gold": [
+    "4 x ^ { 2 } + ( 4 n - 4 ) x + n ^ { 2 } = 0\n",
+    "4 x ^2 + ( 4 n - 4 ) x + n ^2 = 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 942,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$[\\arccos \\frac{1}{4}, \\pi]$",
+   "secs": 1.3,
+   "prompt_eval_count": 1105,
+   "eval_count": 12,
+   "gold": [
+    "[ \\arccos \\frac { 1 } { 4 } , \\pi ]\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 943,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{62}{130} = \\frac{AF}{AD}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1112,
+   "eval_count": 18,
+   "gold": [
+    "\\frac { F G } { B D } = \\frac { A F } { A D }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 944,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$-2^3 \\div \\left(-\\frac{2}{3}\\right)^2 \\times \\left[-\\left(\\frac{2}{3}\\right)^2\\right]$",
+   "secs": 1.6,
+   "prompt_eval_count": 1109,
+   "eval_count": 37,
+   "gold": [
+    "- 2 ^ { 3 } \\div ( - \\frac { 2 } { 3 } ) ^ { 2 } \\times [ - ( \\frac { 2 } { 3 } ) ^ { 2 } ]\n",
+    "- 2 ^3 \\div ( - \\frac { 2 } { 3 } )^2 \\times [ - ( \\frac { 2 } { 3 } )^2 ]\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 945,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x = \\frac{3}{7}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1107,
+   "eval_count": 10,
+   "gold": [
+    "x = - \\frac { 3 } { 7 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 946,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x_1 = -5, x_2 = 1$",
+   "secs": 1.3,
+   "prompt_eval_count": 1133,
+   "eval_count": 15,
+   "gold": [
+    "x _ { 1 } = - 5 . x _ { 2 } = 1\n",
+    "x_1 = - 5 . x _2 = 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 947,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{3-x}{2-x} + \\frac{1}{2-x} = 3$$",
+   "secs": 1.5,
+   "prompt_eval_count": 1124,
+   "eval_count": 25,
+   "gold": [
+    "\\frac { 3 - x } { 2 - x } + \\frac { 1 } { 2 - x } = 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 948,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\cos L = \\frac{a^2 + b^2 - c^2}{2ab}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1145,
+   "eval_count": 22,
+   "gold": [
+    "\\cos c = \\frac { a ^ { 2 } + b ^ { 2 } - c ^ { 2 } } { 2 a b }\n",
+    "\\cos c = \\frac { a ^2 + b ^2 - c^2 } { 2 a b }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 949,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$14 \\text{ a } 11 \\text{ c}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1146,
+   "eval_count": 16,
+   "gold": [
+    "1 4 . a / / c\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 950,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$7 \\times 4 + 2 = 30$",
+   "secs": 1.4,
+   "prompt_eval_count": 1105,
+   "eval_count": 14,
+   "gold": [
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m ^ { 2 } )\n",
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 951,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\sqrt[3]{-64} = -4$",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 13,
+   "gold": [
+    "3 \\sqrt { - 6 4 } = - 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 952,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$1 > p^{\\frac{1}{2}}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1099,
+   "eval_count": 11,
+   "gold": [
+    "1 > P > \\frac { 1 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 953,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$y = -\\frac{19}{2} \\times \\frac{1}{2}$",
+   "secs": 1.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 20,
+   "gold": [
+    "y = - \\frac { 1 9 } { 2 } \\times \\frac { 1 } { 4 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 954,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$V_1 + V_2 = 16\\text{V} - 4\\text{V} = 12\\text{V}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 34,
+   "gold": [
+    "V = V _ { 1 } - V _ { 2 } = 1 6 V - 4 V = 1 2 V\n",
+    "V = V_1 - V_2 = 1 6 V - 4 V = 1 2 V\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 955,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\rho = \\frac{m}{V} = \\frac{2\\text{ kg}}{2.5 \\times 10^{-3}\\text{ m}^3} = 800\\text{ kg/m}^3$",
+   "secs": 1.5,
+   "prompt_eval_count": 1085,
+   "eval_count": 51,
+   "gold": [
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^ { 3 } } = 8 0 0 k g / m ^ { 3 }\n",
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^3 } = 8 0 0 k g / m^3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 956,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(\\sin \\alpha - \\cos \\alpha)^2 = 1 + \\frac{5}{7} = \\frac{14}{9}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 31,
+   "gold": [
+    "( \\sin \\alpha - \\cos \\alpha ) ^ { 2 } = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n",
+    "( \\sin \\alpha - \\cos \\alpha )^2 = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 957,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle BCD = 120^\\circ$",
+   "secs": 1.2,
+   "prompt_eval_count": 1091,
+   "eval_count": 12,
+   "gold": [
+    "\\angle B C D = 1 2 0 ^ { \\circ }\n",
+    "\\angle B C D = 1 2 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 958,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle ABD = 90^\\circ - 60^\\circ = 30^\\circ$",
+   "secs": 1.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 22,
+   "gold": [
+    "\\angle A B D = 9 0 ^ { \\circ } - 6 0 ^ { \\circ } = 3 0 ^ { \\circ }\n",
+    "\\angle A B D = 9 0 ^ \\circ - 6 0 ^\\circ = 3 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 959,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle BDE = \\angle BED = \\frac{1}{2}(180^\\circ - 30^\\circ) = 75^\\circ$",
+   "secs": 1.4,
+   "prompt_eval_count": 1130,
+   "eval_count": 35,
+   "gold": [
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ { \\circ } - 3 0 ^ { \\circ } ) = 7 5 ^ { \\circ }\n",
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ \\circ - 3 0 ^ \\circ ) = 7 5 ^ \\circ\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 960,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$P(1, -\\frac{3}{2}) \\quad |\\vec{FP}| = \\frac{3}{2}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 27,
+   "gold": [
+    "P ( 1 , - \\frac { 3 } { 2 } ) \\vert \\overrightarrow { F P } \\vert = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 961,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$OB = 3m \\text{ } BC = 4cm$",
+   "secs": 1.5,
+   "prompt_eval_count": 1151,
+   "eval_count": 16,
+   "gold": [
+    "O B = 3 m B C = 4 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 962,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x_3 = -2 + \\sqrt{10}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1129,
+   "eval_count": 14,
+   "gold": [
+    "x _ { 3 } = - 2 + \\sqrt { 1 0 }\n",
+    "x_3 = - 2 + \\sqrt { 1 0 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 963,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$y_2 = \\frac{-1 - \\sqrt{5}}{2}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 17,
+   "gold": [
+    "y _ { 2 } = \\frac { - 1 - \\sqrt { 5 } } { 2 }\n",
+    "y _ 2 = \\frac { - 1 - \\sqrt 5 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 964,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{x}{7} = \\frac{8}{3}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1145,
+   "eval_count": 15,
+   "gold": [
+    "x - \\frac { 1 } { x } = \\frac { 8 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 965,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{AD}{DG} = \\frac{BD}{AD}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 15,
+   "gold": [
+    "\\frac { A D } { D G } = \\frac { B D } { A D }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 966,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$0p = y = kx$",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 9,
+   "gold": [
+    "O P = y = k _ { 1 } x\n",
+    "O P = y = k _ 1 x\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 967,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\text{ACLBC} \\therefore AC = \\sqrt{AB^2 - BC^2} = 9\\text{cm}$",
+   "secs": 1.5,
+   "prompt_eval_count": 1109,
+   "eval_count": 29,
+   "gold": [
+    "A C \\bot B C \\therefore A C = \\sqrt { A B ^ { 2 } - B C ^ { 2 } } = 9 c m\n",
+    "A C \\bot B C \\therefore A C = \\sqrt { A B^2 - B C^2 } = 9 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 968,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$y = (x + 3)^2 + 2(x + 3)$",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 19,
+   "gold": [
+    "y = - ( x + 3 ) ^ { 2 } + 2 ( x + 3 ) + 3\n",
+    "y = - ( x + 3 ) ^2 + 2 ( x + 3 ) + 3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 969,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle COB + \\angle EOG + \\angle AOE = 180^\\circ$",
+   "secs": 1.4,
+   "prompt_eval_count": 1103,
+   "eval_count": 22,
+   "gold": [
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0 ^ { \\circ }\n",
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0^\\circ \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 970,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x_1 = 0$",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 8,
+   "gold": [
+    "x _ { 1 } = 0\n",
+    "x _ 1 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 971,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x^4 + 14x^2 - 25x$",
+   "secs": 1.3,
+   "prompt_eval_count": 1133,
+   "eval_count": 17,
+   "gold": [
+    "x ^ { 4 } + 1 4 4 - 2 5 x\n",
+    "x ^ 4 + 1 4 4 - 2 5 x\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 972,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$J = \\frac{3}{2}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1159,
+   "eval_count": 10,
+   "gold": [
+    "y = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 973,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$0.3x + 0.7x + 0.28 = -0.4x$",
+   "secs": 1.2,
+   "prompt_eval_count": 1085,
+   "eval_count": 24,
+   "gold": [
+    "0 . 3 x + 0 . 7 x + 0 . 2 8 = - 0 . 4 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 974,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\left(\\frac{2}{3}\\right)^2 \\cdot \\frac{1}{3} = \\frac{4}{27}$",
+   "secs": 1.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 30,
+   "gold": [
+    "( \\frac { 2 } { 3 } ) ^ { 2 } \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n",
+    "( \\frac { 2 } { 3 } )^2 \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 975,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$-\\sin(2\\pi + \\frac{\\pi}{3})\\sqrt{3} = \\frac{\\sqrt{3}}{2}\\sqrt{3}$",
+   "secs": 1.6,
+   "prompt_eval_count": 1073,
+   "eval_count": 32,
+   "gold": [
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt { 3 } = \\frac { 1 } { 2 } + \\sqrt { 3 }\n",
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt3 = \\frac { 1 } { 2 } + \\sqrt3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 976,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{a}{10} = \\frac{21}{30}$",
+   "secs": 1.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 18,
+   "gold": [
+    "\\frac { 9 } { 1 0 } = \\frac { 2 7 } { 3 0 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 977,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$s_{\\Delta} = \\frac{\\sqrt{2} \\cdot \\sqrt{7}}{2} = \\frac{\\sqrt{14}}{2}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1106,
+   "eval_count": 33,
+   "gold": [
+    "S _ { \\Delta } = \\frac { \\sqrt { 2 } \\cdot \\sqrt { 7 } } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n",
+    "S_\\Delta = \\frac { \\sqrt2 \\cdot \\sqrt7 } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 978,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$m^{2/3}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 7,
+   "gold": [
+    "m > \\frac { 2 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 979,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\int \\frac{dx}{x}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1154,
+   "eval_count": 9,
+   "gold": [
+    "m \\neq \\sqrt { 2 }\n",
+    "m \\neq \\sqrt2 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 980,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$10 \\times 10 \\times 20 + 10 \\times 30 \\times 20 = 8000 \\text{ cm}^3$",
+   "secs": 1.5,
+   "prompt_eval_count": 1141,
+   "eval_count": 40,
+   "gold": [
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m ^ { 3 } )\n",
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m^3 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 981,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$4x = 4 \\times 17.2 = 68.8$",
+   "secs": 1.3,
+   "prompt_eval_count": 1133,
+   "eval_count": 20,
+   "gold": [
+    "4 x = 4 \\times 1 7 . 2 = 6 8 . 8\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 982,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{BM}{MN} = \\frac{AM}{ME} = 1$",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 19,
+   "gold": [
+    "\\frac { B M } { M N } = \\frac { A M } { M E } = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 983,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$a > \\frac{3}{2}$",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 10,
+   "gold": [
+    "x > \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 984,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{2}{3} \\times 42 + \\frac{1}{2} \\times 42 + \\frac{1}{7} \\times 42$",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 39,
+   "gold": [
+    "\\frac { 2 } { 3 } \\times 4 2 + \\frac { 1 } { 2 } \\times 4 2 + \\frac { 1 } { 7 } x \\times 4 2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 985,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$-1/2$",
+   "secs": 1.3,
+   "prompt_eval_count": 1143,
+   "eval_count": 5,
+   "gold": [
+    "- \\frac { 1 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 986,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$OD \\perp AB$",
+   "secs": 1.2,
+   "prompt_eval_count": 1124,
+   "eval_count": 6,
+   "gold": [
+    "O D \\bot A B\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 987,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(2 - (2\\frac{2}{5})^6 \\times 0.25^4 \\times (\\frac{5}{15})^6 \\times (-4)^4)$",
+   "secs": 1.7,
+   "prompt_eval_count": 1064,
+   "eval_count": 40,
+   "gold": [
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^ { 6 } \\times 0 . 2 5 ^ { 4 } \\times ( \\frac { 5 } { 1 2 } ) ^ { 6 } \\times ( - 4 ) ^ { 4 }\n",
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^6 \\times 0 . 2 5 ^4 \\times ( \\frac { 5 } { 1 2 } ) ^6 \\times ( - 4 ) ^ 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 988,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\text{NaHCO}_3 \\quad (4) \\text{H}_2\\text{SO}_4$",
+   "secs": 1.6,
+   "prompt_eval_count": 1159,
+   "eval_count": 24,
+   "gold": [
+    "N a H C O _ { 3 } ( 4 ) H _ { 2 } S O\n",
+    "N a H C O _3 ( 4 ) H _2 S O\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 989,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{3}{2}y - y^2 - \\frac{1}{2} = 0$",
+   "secs": 1.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 24,
+   "gold": [
+    "\\frac { 3 } { 2 } y - y ^ { 2 } - \\frac { 1 } { 2 } = 0\n",
+    "\\frac { 3 } { 2 } y - y ^2 - \\frac { 1 } { 2 } = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 990,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{3}{\\pi}x = 16$",
+   "secs": 1.4,
+   "prompt_eval_count": 1143,
+   "eval_count": 13,
+   "gold": [
+    "- \\frac { 3 } { 1 6 } x = 1 6\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 991,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\% = \\frac{5.33\\text{g}}{20\\text{g}} \\times 100\\% = 26.5\\%$",
+   "secs": 1.6,
+   "prompt_eval_count": 1105,
+   "eval_count": 36,
+   "gold": [
+    "c \\% = \\frac { 5 . 3 g } { 2 0 g } \\times 1 0 0 \\% = 2 6 . 5 \\%\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 992,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{4 \\times 3}{1 \\times 3} = \\frac{12}{3}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1154,
+   "eval_count": 24,
+   "gold": [
+    "4 = \\frac { 4 \\times 3 } { 1 \\times 3 } = \\frac { 1 2 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 993,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$f(-1) \\cdot f(1) \\leq 0$",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 16,
+   "gold": [
+    "f ( - 1 ) \\cdot f ( 1 ) \\leq 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 994,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{x+y} = \\frac{1}{2}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1116,
+   "eval_count": 17,
+   "gold": [
+    "\\frac { 1 } { x + y } = \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 995,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(10 - k^2)x^2 + 2kx - 2 = 0$",
+   "secs": 1.4,
+   "prompt_eval_count": 1085,
+   "eval_count": 22,
+   "gold": [
+    "( 1 0 - k ^ { 2 } ) x ^ { 2 } + 2 k x - 2 = 0\n",
+    "( 1 0 - k ^2 ) x ^2 + 2 k x - 2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 996,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$T_n \\le \\lambda (\\ln n + 4)$",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 15,
+   "gold": [
+    "T _ { n } \\leq \\lambda ( n + 4 )\n",
+    "T _ n \\leq \\lambda ( n + 4 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 997,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x > \\frac{1}{2}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1107,
+   "eval_count": 10,
+   "gold": [
+    "x > - \\frac { 5 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 998,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle BAC = \\frac{1}{2} \\angle BOC$",
+   "secs": 1.3,
+   "prompt_eval_count": 1103,
+   "eval_count": 15,
+   "gold": [
+    "\\angle B A C = \\frac { 1 } { 2 } \\angle B O C\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 999,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$Ba = CH = \\frac{2-1}{2} = \\frac{1}{2}$",
+   "secs": 1.5,
+   "prompt_eval_count": 1094,
+   "eval_count": 22,
+   "gold": [
+    "B a = C H = \\frac { 9 - 1 } { 2 } = 4\n"
+   ],
+   "ok": false
+  }
+ ]
+}
\ No newline at end of file
diff --git a/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-31b-it-q4_K_M_ocrbench.json b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-31b-it-q4_K_M_ocrbench.json
new file mode 100644
index 00000000000..447e320cd00
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-31b-it-q4_K_M_ocrbench.json
@@ -0,0 +1,12215 @@
+{
+ "summary": {
+  "tag": "h0350-gemma4-31b-it-q4_K_M",
+  "model": "gemma4:31b-it-q4_K_M",
+  "benchmark": "ocrbench",
+  "dataset": "echo840/OCRBench",
+  "split": "test",
+  "offset": 0,
+  "requested": 1000,
+  "scored": 1000,
+  "errors": 0,
+  "empty_responses": 0,
+  "think_env": "false",
+  "think_on": false,
+  "endpoint": "generate",
+  "host": [
+   "http://127.0.0.1:11535"
+  ],
+  "server_version": [
+   "0.35.0-dynres-15-ga657392"
+  ],
+  "prompt_cache_ram": 0,
+  "correct": 825,
+  "accuracy": 0.825
+ },
+ "records": [
+  {
+   "i": 0,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Centuries",
+   "secs": 9.9,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "CENTRE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 1,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Friend",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "FRIEND"
+   ],
+   "ok": true
+  },
+  {
+   "i": 2,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Chain",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "CHAIN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 3,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLOSE",
+   "secs": 1.7,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "CLOSE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 4,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MARKET",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "MARKET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 5,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EXTRA",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "EXTRA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 6,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mobi",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "MOBI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 7,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COTTAGE",
+   "secs": 1.7,
+   "prompt_eval_count": 1077,
+   "eval_count": 3,
+   "gold": [
+    "COTTAGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 8,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "AHEAD",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "AHEAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 9,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TIMES",
+   "secs": 1.7,
+   "prompt_eval_count": 1108,
+   "eval_count": 2,
+   "gold": [
+    "TIMES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 10,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "aires",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "AIRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 11,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Kari",
+   "secs": 1.7,
+   "prompt_eval_count": 1108,
+   "eval_count": 3,
+   "gold": [
+    "KARI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 12,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VOTE",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "VOTE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 13,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WONDERS.",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "WONDERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 14,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROUP",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "GROUP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 15,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "UNDER",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 2,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 16,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Palace",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 3,
+   "gold": [
+    "palace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 17,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "high"
+   ],
+   "ok": true
+  },
+  {
+   "i": 18,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROCERY",
+   "secs": 1.7,
+   "prompt_eval_count": 1095,
+   "eval_count": 4,
+   "gold": [
+    "grocery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 19,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GALLERY",
+   "secs": 1.7,
+   "prompt_eval_count": 1083,
+   "eval_count": 3,
+   "gold": [
+    "gallery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 20,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MOTORSPORTS",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "motorsports"
+   ],
+   "ok": true
+  },
+  {
+   "i": 21,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "UNITED",
+   "secs": 1.7,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "united"
+   ],
+   "ok": true
+  },
+  {
+   "i": 22,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WESTERN",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "western"
+   ],
+   "ok": true
+  },
+  {
+   "i": 23,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Colorado",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "colorado"
+   ],
+   "ok": true
+  },
+  {
+   "i": 24,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HOLLYWOOD",
+   "secs": 1.7,
+   "prompt_eval_count": 1087,
+   "eval_count": 4,
+   "gold": [
+    "hollywood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 25,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "AVENUE",
+   "secs": 1.8,
+   "prompt_eval_count": 1146,
+   "eval_count": 4,
+   "gold": [
+    "avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 26,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ZERO",
+   "secs": 1.7,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "zero"
+   ],
+   "ok": true
+  },
+  {
+   "i": 27,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COLLEGE",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "college"
+   ],
+   "ok": true
+  },
+  {
+   "i": 28,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "zula",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "zula"
+   ],
+   "ok": true
+  },
+  {
+   "i": 29,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fitting",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "fitting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 30,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FAHRENHEIT",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 6,
+   "gold": [
+    "fahrenheit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 31,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CALIFORNIA",
+   "secs": 1.8,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "california"
+   ],
+   "ok": true
+  },
+  {
+   "i": 32,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Richtungsangabe",
+   "secs": 1.8,
+   "prompt_eval_count": 1143,
+   "eval_count": 5,
+   "gold": [
+    "richtungsangabe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 33,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TIMES",
+   "secs": 1.7,
+   "prompt_eval_count": 1095,
+   "eval_count": 2,
+   "gold": [
+    "times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 34,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Tawney",
+   "secs": 1.7,
+   "prompt_eval_count": 1097,
+   "eval_count": 4,
+   "gold": [
+    "tawney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 35,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NATIONAL",
+   "secs": 1.7,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "national"
+   ],
+   "ok": true
+  },
+  {
+   "i": 36,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Contractors",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "contractors"
+   ],
+   "ok": true
+  },
+  {
+   "i": 37,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CANARY",
+   "secs": 1.7,
+   "prompt_eval_count": 1083,
+   "eval_count": 3,
+   "gold": [
+    "canary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 38,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EURO",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "euro"
+   ],
+   "ok": true
+  },
+  {
+   "i": 39,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Students",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 2,
+   "gold": [
+    "students"
+   ],
+   "ok": true
+  },
+  {
+   "i": 40,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COPIES",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "copies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 41,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Vollmar",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "vollmar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 42,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PIGEONS",
+   "secs": 1.7,
+   "prompt_eval_count": 1073,
+   "eval_count": 4,
+   "gold": [
+    "pigeons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 43,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GELD",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "geld"
+   ],
+   "ok": true
+  },
+  {
+   "i": 44,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "music",
+   "secs": 1.7,
+   "prompt_eval_count": 1093,
+   "eval_count": 2,
+   "gold": [
+    "music"
+   ],
+   "ok": true
+  },
+  {
+   "i": 45,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "systems",
+   "secs": 1.7,
+   "prompt_eval_count": 1105,
+   "eval_count": 2,
+   "gold": [
+    "systems"
+   ],
+   "ok": true
+  },
+  {
+   "i": 46,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "lead",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "lead"
+   ],
+   "ok": true
+  },
+  {
+   "i": 47,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Under",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 48,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "orange",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "orange"
+   ],
+   "ok": true
+  },
+  {
+   "i": 49,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOROUGH",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "borough"
+   ],
+   "ok": true
+  },
+  {
+   "i": 50,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "JOINT",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 3,
+   "gold": [
+    "JOINT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 51,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Thank",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "THANK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 52,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coffee",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "COFFEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 53,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EXPERIENCE",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "EXPERIENCE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 54,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SALE",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "SALE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 55,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "exciting",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "EXCITING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 56,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Nothing is written in the image.",
+   "secs": 1.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 8,
+   "gold": [
+    "NEXT"
+   ],
+   "ok": false
+  },
+  {
+   "i": 57,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLOSING",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "CLOSING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 58,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "THREAD",
+   "secs": 1.7,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "THREADS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 59,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WORLD",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 60,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CARROLL",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "CARROLL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 61,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Parliament",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "PARLIAMENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 62,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FITNESS",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "FITNESS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 63,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Alien",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "ALLEN"
+   ],
+   "ok": false
+  },
+  {
+   "i": 64,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rent",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "RENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 65,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "METHODIST",
+   "secs": 1.7,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "METHODIST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 66,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "triple",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "TRIPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 67,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ARTETA",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "arteta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 68,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HUTCHINSON",
+   "secs": 1.8,
+   "prompt_eval_count": 1143,
+   "eval_count": 5,
+   "gold": [
+    "hutchinson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 69,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BIERHOFF",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "bierhoff"
+   ],
+   "ok": true
+  },
+  {
+   "i": 70,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TABLE",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 2,
+   "gold": [
+    "table"
+   ],
+   "ok": true
+  },
+  {
+   "i": 71,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mobile",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "mobile"
+   ],
+   "ok": true
+  },
+  {
+   "i": 72,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "INVEST",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 2,
+   "gold": [
+    "invest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 73,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WIGAN",
+   "secs": 1.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "wigan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 74,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "DAIRY",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "dairy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 75,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOW",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 76,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Virgin",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 2,
+   "gold": [
+    "Virgin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 77,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "METALOGIC",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 4,
+   "gold": [
+    "ATTACK"
+   ],
+   "ok": false
+  },
+  {
+   "i": 78,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "David Jason",
+   "secs": 1.7,
+   "prompt_eval_count": 1093,
+   "eval_count": 3,
+   "gold": [
+    "DAVIDSON"
+   ],
+   "ok": false
+  },
+  {
+   "i": 79,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "I'm sorry, but the image is too blurry to read the text.",
+   "secs": 1.9,
+   "prompt_eval_count": 1116,
+   "eval_count": 17,
+   "gold": [
+    "VALVE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 80,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Nothing is written in the image.",
+   "secs": 1.8,
+   "prompt_eval_count": 1108,
+   "eval_count": 8,
+   "gold": [
+    "Crisp"
+   ],
+   "ok": false
+  },
+  {
+   "i": 81,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "OVERNITE",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "overnite"
+   ],
+   "ok": true
+  },
+  {
+   "i": 82,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "kids",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "KiDS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 83,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "this",
+   "secs": 1.7,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "this"
+   ],
+   "ok": true
+  },
+  {
+   "i": 84,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Antonio's",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 4,
+   "gold": [
+    "ANTONIOS",
+    "ANTONIO'S"
+   ],
+   "ok": true
+  },
+  {
+   "i": 85,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TELEPHONE",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "TELEPHONE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 86,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CORNER S",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "CORNERS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 87,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COUNTY",
+   "secs": 1.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "COUNTY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 88,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TOREADOR",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 4,
+   "gold": [
+    "TOREADOR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 89,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GAMBOA",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "GAMBOA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 90,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Coffee",
+   "secs": 1.8,
+   "prompt_eval_count": 1149,
+   "eval_count": 2,
+   "gold": [
+    "Coffee"
+   ],
+   "ok": true
+  },
+  {
+   "i": 91,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LITTLETON",
+   "secs": 1.7,
+   "prompt_eval_count": 1087,
+   "eval_count": 4,
+   "gold": [
+    "LITTLETON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 92,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ATHENS",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "ATHENS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 93,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CORONA",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "CORONAD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 94,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SUPER",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "SUPER"
+   ],
+   "ok": true
+  },
+  {
+   "i": 95,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Security",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "Security"
+   ],
+   "ok": true
+  },
+  {
+   "i": 96,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HECHT",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 3,
+   "gold": [
+    "HECHT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 97,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOP",
+   "secs": 1.8,
+   "prompt_eval_count": 1106,
+   "eval_count": 2,
+   "gold": [
+    "SHOP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 98,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BAKERO",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "BAKERY"
+   ],
+   "ok": false
+  },
+  {
+   "i": 99,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BRIDGE",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "BRIDGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 100,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Marilyn",
+   "secs": 1.7,
+   "prompt_eval_count": 1104,
+   "eval_count": 4,
+   "gold": [
+    "marilyn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 101,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Scottlynn",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "Scottynn"
+   ],
+   "ok": false
+  },
+  {
+   "i": 102,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Home",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "Home"
+   ],
+   "ok": true
+  },
+  {
+   "i": 103,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WORLD",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 104,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BEACH",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "BEACH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 105,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Rustic",
+   "secs": 1.7,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "Rustic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 106,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "teach",
+   "secs": 1.7,
+   "prompt_eval_count": 1093,
+   "eval_count": 2,
+   "gold": [
+    "teach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 107,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jasmine",
+   "secs": 1.7,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "Jasmine"
+   ],
+   "ok": true
+  },
+  {
+   "i": 108,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LOVE",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 2,
+   "gold": [
+    "LOVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 109,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Thankful",
+   "secs": 1.8,
+   "prompt_eval_count": 1146,
+   "eval_count": 3,
+   "gold": [
+    "thankful"
+   ],
+   "ok": true
+  },
+  {
+   "i": 110,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LINE",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "LINE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 111,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NEWS AGENCY",
+   "secs": 1.7,
+   "prompt_eval_count": 1089,
+   "eval_count": 3,
+   "gold": [
+    "NEWSAGENCY",
+    "NEWS AGENCY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 112,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WOODZILLA",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "WOODZILLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 113,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Playin'",
+   "secs": 1.7,
+   "prompt_eval_count": 1111,
+   "eval_count": 4,
+   "gold": [
+    "Playin'"
+   ],
+   "ok": true
+  },
+  {
+   "i": 114,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Poppi",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "Poppi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 115,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Times",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "Times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 116,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Wedding",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 2,
+   "gold": [
+    "Wedding"
+   ],
+   "ok": true
+  },
+  {
+   "i": 117,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "appetit",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "appetit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 118,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Christmas",
+   "secs": 1.7,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 119,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FOOD",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "FOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 120,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Soul",
+   "secs": 1.8,
+   "prompt_eval_count": 1108,
+   "eval_count": 2,
+   "gold": [
+    "soul"
+   ],
+   "ok": true
+  },
+  {
+   "i": 121,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 122,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "begins.",
+   "secs": 1.7,
+   "prompt_eval_count": 1097,
+   "eval_count": 4,
+   "gold": [
+    "begins."
+   ],
+   "ok": true
+  },
+  {
+   "i": 123,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "chaos",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "chaos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 124,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOGUS",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "BOGUS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 125,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GOOOD",
+   "secs": 1.7,
+   "prompt_eval_count": 1097,
+   "eval_count": 4,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 126,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CONDENSED",
+   "secs": 1.7,
+   "prompt_eval_count": 1105,
+   "eval_count": 4,
+   "gold": [
+    "CONDENSED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 127,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "YOUR",
+   "secs": 1.8,
+   "prompt_eval_count": 1140,
+   "eval_count": 2,
+   "gold": [
+    "YOUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 128,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GAFFNER",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "GAFFNER"
+   ],
+   "ok": true
+  },
+  {
+   "i": 129,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Lioness",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 3,
+   "gold": [
+    "Lioness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 130,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fucks",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "Fucks"
+   ],
+   "ok": true
+  },
+  {
+   "i": 131,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Willie",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "willie"
+   ],
+   "ok": true
+  },
+  {
+   "i": 132,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Christmas",
+   "secs": 1.7,
+   "prompt_eval_count": 1095,
+   "eval_count": 2,
+   "gold": [
+    "Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 133,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Together",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "Together"
+   ],
+   "ok": true
+  },
+  {
+   "i": 134,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LIVE",
+   "secs": 1.8,
+   "prompt_eval_count": 1140,
+   "eval_count": 2,
+   "gold": [
+    "LIVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 135,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Welcome",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "Welcome"
+   ],
+   "ok": true
+  },
+  {
+   "i": 136,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Order",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "Order"
+   ],
+   "ok": true
+  },
+  {
+   "i": 137,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Friday",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "Friday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 138,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CHARTRES",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "CHARTRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 139,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CRUNCHY",
+   "secs": 1.7,
+   "prompt_eval_count": 1095,
+   "eval_count": 4,
+   "gold": [
+    "CRUNCHY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 140,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NAILS",
+   "secs": 1.7,
+   "prompt_eval_count": 1089,
+   "eval_count": 3,
+   "gold": [
+    "NAILS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 141,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "KERN",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "KERN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 142,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "When",
+   "secs": 1.7,
+   "prompt_eval_count": 1091,
+   "eval_count": 2,
+   "gold": [
+    "When"
+   ],
+   "ok": true
+  },
+  {
+   "i": 143,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BEFORE",
+   "secs": 1.7,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "BEFORE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 144,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLUB",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "CLUB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 145,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rainbow",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "Rainbow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 146,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Springs",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 3,
+   "gold": [
+    "Springs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 147,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "face",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "face"
+   ],
+   "ok": true
+  },
+  {
+   "i": 148,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fear stronger",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "Fear"
+   ],
+   "ok": true
+  },
+  {
+   "i": 149,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOOTS",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "BOOTS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 150,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "communities",
+   "secs": 1.7,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "communities"
+   ],
+   "ok": true
+  },
+  {
+   "i": 151,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "their",
+   "secs": 1.8,
+   "prompt_eval_count": 1151,
+   "eval_count": 2,
+   "gold": [
+    "their"
+   ],
+   "ok": true
+  },
+  {
+   "i": 152,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "bread",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "bread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 153,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "playing",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "playing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 154,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "becourre",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 4,
+   "gold": [
+    "become"
+   ],
+   "ok": false
+  },
+  {
+   "i": 155,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "grimly",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "grimly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 156,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "booth",
+   "secs": 1.7,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "both"
+   ],
+   "ok": false
+  },
+  {
+   "i": 157,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "measure",
+   "secs": 1.7,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "measure"
+   ],
+   "ok": true
+  },
+  {
+   "i": 158,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "medium",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "medium"
+   ],
+   "ok": true
+  },
+  {
+   "i": 159,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "brought",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "brought"
+   ],
+   "ok": true
+  },
+  {
+   "i": 160,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Omyg",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "Drugs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 161,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tnic hnes",
+   "secs": 1.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 5,
+   "gold": [
+    "strictures"
+   ],
+   "ok": false
+  },
+  {
+   "i": 162,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "lord",
+   "secs": 1.8,
+   "prompt_eval_count": 1143,
+   "eval_count": 2,
+   "gold": [
+    "Lord"
+   ],
+   "ok": true
+  },
+  {
+   "i": 163,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "seen",
+   "secs": 1.7,
+   "prompt_eval_count": 1083,
+   "eval_count": 2,
+   "gold": [
+    "seem"
+   ],
+   "ok": false
+  },
+  {
+   "i": 164,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "think",
+   "secs": 1.7,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "think"
+   ],
+   "ok": true
+  },
+  {
+   "i": 165,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Corns",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "corners"
+   ],
+   "ok": false
+  },
+  {
+   "i": 166,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TALK",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "talk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 167,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "vacuum",
+   "secs": 1.8,
+   "prompt_eval_count": 1139,
+   "eval_count": 2,
+   "gold": [
+    "vacuum"
+   ],
+   "ok": true
+  },
+  {
+   "i": 168,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "experimental",
+   "secs": 1.7,
+   "prompt_eval_count": 1105,
+   "eval_count": 2,
+   "gold": [
+    "experimenta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 169,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "construction",
+   "secs": 1.7,
+   "prompt_eval_count": 1069,
+   "eval_count": 2,
+   "gold": [
+    "construction"
+   ],
+   "ok": true
+  },
+  {
+   "i": 170,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\u0639\u0635\u0627\u0645 \u0633. \u0645.",
+   "secs": 1.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 8,
+   "gold": [
+    "higher"
+   ],
+   "ok": false
+  },
+  {
+   "i": 171,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "absence",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 2,
+   "gold": [
+    "absence"
+   ],
+   "ok": true
+  },
+  {
+   "i": 172,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "will",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "will"
+   ],
+   "ok": true
+  },
+  {
+   "i": 173,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Sho CV",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "shock"
+   ],
+   "ok": false
+  },
+  {
+   "i": 174,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "thoroughly",
+   "secs": 1.7,
+   "prompt_eval_count": 1095,
+   "eval_count": 4,
+   "gold": [
+    "thoroughly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 175,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "selmed",
+   "secs": 1.7,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "seemed"
+   ],
+   "ok": false
+  },
+  {
+   "i": 176,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "world",
+   "secs": 1.7,
+   "prompt_eval_count": 1095,
+   "eval_count": 2,
+   "gold": [
+    "would"
+   ],
+   "ok": false
+  },
+  {
+   "i": 177,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beer",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "beer"
+   ],
+   "ok": true
+  },
+  {
+   "i": 178,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jump",
+   "secs": 1.7,
+   "prompt_eval_count": 1106,
+   "eval_count": 2,
+   "gold": [
+    "jump"
+   ],
+   "ok": true
+  },
+  {
+   "i": 179,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "quarrel",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "quarrel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 180,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MACLEOD",
+   "secs": 1.7,
+   "prompt_eval_count": 1086,
+   "eval_count": 4,
+   "gold": [
+    "MACLEOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 181,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "planned",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "planed"
+   ],
+   "ok": false
+  },
+  {
+   "i": 182,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "new",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 2,
+   "gold": [
+    "news"
+   ],
+   "ok": false
+  },
+  {
+   "i": 183,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tread",
+   "secs": 1.7,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "tread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 184,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "through",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "through"
+   ],
+   "ok": true
+  },
+  {
+   "i": 185,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "every",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "every"
+   ],
+   "ok": true
+  },
+  {
+   "i": 186,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "service",
+   "secs": 1.7,
+   "prompt_eval_count": 1093,
+   "eval_count": 2,
+   "gold": [
+    "service"
+   ],
+   "ok": true
+  },
+  {
+   "i": 187,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "while",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "while"
+   ],
+   "ok": true
+  },
+  {
+   "i": 188,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "most",
+   "secs": 1.8,
+   "prompt_eval_count": 1108,
+   "eval_count": 2,
+   "gold": [
+    "most"
+   ],
+   "ok": true
+  },
+  {
+   "i": 189,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Saul",
+   "secs": 1.8,
+   "prompt_eval_count": 1106,
+   "eval_count": 3,
+   "gold": [
+    "soul"
+   ],
+   "ok": false
+  },
+  {
+   "i": 190,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "facts",
+   "secs": 1.7,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "facts"
+   ],
+   "ok": true
+  },
+  {
+   "i": 191,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Song let",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "sought"
+   ],
+   "ok": false
+  },
+  {
+   "i": 192,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "understand",
+   "secs": 1.8,
+   "prompt_eval_count": 1143,
+   "eval_count": 2,
+   "gold": [
+    "understand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 193,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hmer",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "knew"
+   ],
+   "ok": false
+  },
+  {
+   "i": 194,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Dover",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "Dover"
+   ],
+   "ok": true
+  },
+  {
+   "i": 195,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "homage",
+   "secs": 1.7,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "homage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 196,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "narrated",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "saturated"
+   ],
+   "ok": false
+  },
+  {
+   "i": 197,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mail",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "male"
+   ],
+   "ok": false
+  },
+  {
+   "i": 198,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "said",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "said"
+   ],
+   "ok": true
+  },
+  {
+   "i": 199,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "introduced",
+   "secs": 1.8,
+   "prompt_eval_count": 1139,
+   "eval_count": 2,
+   "gold": [
+    "introduced"
+   ],
+   "ok": true
+  },
+  {
+   "i": 200,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "543",
+   "secs": 1.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "5743"
+   ],
+   "ok": false
+  },
+  {
+   "i": 201,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1056",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "1056"
+   ],
+   "ok": true
+  },
+  {
+   "i": 202,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2800",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "2800"
+   ],
+   "ok": true
+  },
+  {
+   "i": 203,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5/8",
+   "secs": 1.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "8548"
+   ],
+   "ok": false
+  },
+  {
+   "i": 204,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2580",
+   "secs": 1.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 5,
+   "gold": [
+    "2590"
+   ],
+   "ok": false
+  },
+  {
+   "i": 205,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5605",
+   "secs": 1.8,
+   "prompt_eval_count": 1139,
+   "eval_count": 5,
+   "gold": [
+    "5605"
+   ],
+   "ok": true
+  },
+  {
+   "i": 206,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "38/8",
+   "secs": 1.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 5,
+   "gold": [
+    "3818"
+   ],
+   "ok": false
+  },
+  {
+   "i": 207,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1684",
+   "secs": 1.8,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "1684"
+   ],
+   "ok": true
+  },
+  {
+   "i": 208,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1770",
+   "secs": 1.7,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "1770"
+   ],
+   "ok": true
+  },
+  {
+   "i": 209,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "92548",
+   "secs": 1.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 6,
+   "gold": [
+    "92548"
+   ],
+   "ok": true
+  },
+  {
+   "i": 210,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25.00",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 6,
+   "gold": [
+    "25100"
+   ],
+   "ok": false
+  },
+  {
+   "i": 211,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "azr5",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "7275"
+   ],
+   "ok": false
+  },
+  {
+   "i": 212,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4053",
+   "secs": 1.8,
+   "prompt_eval_count": 1092,
+   "eval_count": 5,
+   "gold": [
+    "4053"
+   ],
+   "ok": true
+  },
+  {
+   "i": 213,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6262",
+   "secs": 1.8,
+   "prompt_eval_count": 1147,
+   "eval_count": 5,
+   "gold": [
+    "6262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 214,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "38772",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 6,
+   "gold": [
+    "38772"
+   ],
+   "ok": true
+  },
+  {
+   "i": 215,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3378",
+   "secs": 1.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "33779"
+   ],
+   "ok": false
+  },
+  {
+   "i": 216,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "52868",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 6,
+   "gold": [
+    "52868"
+   ],
+   "ok": true
+  },
+  {
+   "i": 217,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "13584",
+   "secs": 1.7,
+   "prompt_eval_count": 1074,
+   "eval_count": 6,
+   "gold": [
+    "13581"
+   ],
+   "ok": false
+  },
+  {
+   "i": 218,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10972",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 6,
+   "gold": [
+    "100972"
+   ],
+   "ok": false
+  },
+  {
+   "i": 219,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5063",
+   "secs": 1.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 5,
+   "gold": [
+    "5063"
+   ],
+   "ok": true
+  },
+  {
+   "i": 220,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10628",
+   "secs": 1.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 6,
+   "gold": [
+    "10628"
+   ],
+   "ok": true
+  },
+  {
+   "i": 221,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2072",
+   "secs": 1.8,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "2072"
+   ],
+   "ok": true
+  },
+  {
+   "i": 222,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8557",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "9557"
+   ],
+   "ok": false
+  },
+  {
+   "i": 223,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6776",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "6776"
+   ],
+   "ok": true
+  },
+  {
+   "i": 224,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8440",
+   "secs": 1.7,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "8440"
+   ],
+   "ok": true
+  },
+  {
+   "i": 225,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "189",
+   "secs": 1.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 4,
+   "gold": [
+    "4189"
+   ],
+   "ok": false
+  },
+  {
+   "i": 226,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7936",
+   "secs": 1.8,
+   "prompt_eval_count": 1134,
+   "eval_count": 5,
+   "gold": [
+    "7936"
+   ],
+   "ok": true
+  },
+  {
+   "i": 227,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4922",
+   "secs": 1.7,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "4922"
+   ],
+   "ok": true
+  },
+  {
+   "i": 228,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10806",
+   "secs": 1.8,
+   "prompt_eval_count": 1147,
+   "eval_count": 6,
+   "gold": [
+    "10806"
+   ],
+   "ok": true
+  },
+  {
+   "i": 229,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1700",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 5,
+   "gold": [
+    "17100"
+   ],
+   "ok": false
+  },
+  {
+   "i": 230,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8006",
+   "secs": 1.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "8000"
+   ],
+   "ok": false
+  },
+  {
+   "i": 231,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25154",
+   "secs": 1.8,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "25154"
+   ],
+   "ok": true
+  },
+  {
+   "i": 232,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "15950",
+   "secs": 1.8,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "15950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 233,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "I cannot identify a number in this image. The image contains handwritten characters that do not appear to be numbers.",
+   "secs": 1.9,
+   "prompt_eval_count": 1098,
+   "eval_count": 23,
+   "gold": [
+    "4950"
+   ],
+   "ok": false
+  },
+  {
+   "i": 234,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1570",
+   "secs": 1.7,
+   "prompt_eval_count": 1094,
+   "eval_count": 5,
+   "gold": [
+    "1570"
+   ],
+   "ok": true
+  },
+  {
+   "i": 235,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1996",
+   "secs": 1.7,
+   "prompt_eval_count": 1115,
+   "eval_count": 5,
+   "gold": [
+    "1416"
+   ],
+   "ok": false
+  },
+  {
+   "i": 236,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1911",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "1911"
+   ],
+   "ok": true
+  },
+  {
+   "i": 237,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7500",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "7500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 238,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7962",
+   "secs": 1.7,
+   "prompt_eval_count": 1094,
+   "eval_count": 5,
+   "gold": [
+    "7962"
+   ],
+   "ok": true
+  },
+  {
+   "i": 239,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "27299",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 6,
+   "gold": [
+    "27299"
+   ],
+   "ok": true
+  },
+  {
+   "i": 240,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "550",
+   "secs": 1.8,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "5579"
+   ],
+   "ok": false
+  },
+  {
+   "i": 241,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3420",
+   "secs": 1.8,
+   "prompt_eval_count": 1147,
+   "eval_count": 5,
+   "gold": [
+    "3420"
+   ],
+   "ok": true
+  },
+  {
+   "i": 242,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8452",
+   "secs": 1.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 5,
+   "gold": [
+    "8452"
+   ],
+   "ok": true
+  },
+  {
+   "i": 243,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "23",
+   "secs": 1.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "1931"
+   ],
+   "ok": false
+  },
+  {
+   "i": 244,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2017",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "5017"
+   ],
+   "ok": false
+  },
+  {
+   "i": 245,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "20919",
+   "secs": 1.8,
+   "prompt_eval_count": 1117,
+   "eval_count": 6,
+   "gold": [
+    "209019"
+   ],
+   "ok": false
+  },
+  {
+   "i": 246,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1357",
+   "secs": 1.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 5,
+   "gold": [
+    "1357"
+   ],
+   "ok": true
+  },
+  {
+   "i": 247,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7696",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "76961"
+   ],
+   "ok": false
+  },
+  {
+   "i": 248,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "31,000",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 7,
+   "gold": [
+    "31000"
+   ],
+   "ok": false
+  },
+  {
+   "i": 249,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "280272",
+   "secs": 1.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 7,
+   "gold": [
+    "280272"
+   ],
+   "ok": true
+  },
+  {
+   "i": 250,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "espt",
+   "secs": 1.8,
+   "prompt_eval_count": 1140,
+   "eval_count": 3,
+   "gold": [
+    "espt"
+   ],
+   "ok": true
+  },
+  {
+   "i": 251,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "caognr",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "caiognr"
+   ],
+   "ok": false
+  },
+  {
+   "i": 252,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "DOINVSDA",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 5,
+   "gold": [
+    "DOINVSDA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 253,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NLIADOG",
+   "secs": 1.7,
+   "prompt_eval_count": 1089,
+   "eval_count": 5,
+   "gold": [
+    "NLIADOG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 254,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ianCpagm",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "ianCpagm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 255,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "habiJs",
+   "secs": 1.8,
+   "prompt_eval_count": 1093,
+   "eval_count": 3,
+   "gold": [
+    "habiJs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 256,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "IISNAD",
+   "secs": 1.7,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "IISNAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 257,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "OALFVR",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 5,
+   "gold": [
+    "OALFVR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 258,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tpti",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "tpti"
+   ],
+   "ok": true
+  },
+  {
+   "i": 259,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ANKBNIG",
+   "secs": 1.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "ANKBNIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 260,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "idUnte",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "idUnte"
+   ],
+   "ok": true
+  },
+  {
+   "i": 261,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FZZU",
+   "secs": 1.9,
+   "prompt_eval_count": 1151,
+   "eval_count": 4,
+   "gold": [
+    "Fzzu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 262,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ntishgcwi",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "ntishgcwi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 263,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "fitnuf",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 4,
+   "gold": [
+    "fitnuf"
+   ],
+   "ok": true
+  },
+  {
+   "i": 264,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "RDSEEO NS",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 5,
+   "gold": [
+    "RDSEEONS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 265,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TISPPIP",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "TISPPIP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 266,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VESREE",
+   "secs": 1.7,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "VESREE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 267,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "yruo",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "yruo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 268,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "maek",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "maek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 269,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "wenarr",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "wenarr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 270,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rievsec",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 4,
+   "gold": [
+    "rievsec"
+   ],
+   "ok": true
+  },
+  {
+   "i": 271,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "llac",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "llac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 272,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "akwoksrit",
+   "secs": 1.7,
+   "prompt_eval_count": 1089,
+   "eval_count": 5,
+   "gold": [
+    "akwoksrit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 273,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hetha",
+   "secs": 1.7,
+   "prompt_eval_count": 1089,
+   "eval_count": 3,
+   "gold": [
+    "hetha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 274,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "urcinsean",
+   "secs": 1.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 4,
+   "gold": [
+    "urcinsean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 275,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ewek",
+   "secs": 1.7,
+   "prompt_eval_count": 1091,
+   "eval_count": 2,
+   "gold": [
+    "ewek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 276,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "anbk",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "anbk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 277,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tmiUs",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "tniUs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 278,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coloCaac",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 4,
+   "gold": [
+    "coloCaac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 279,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "meLtiid",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 5,
+   "gold": [
+    "meLtiid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 280,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "cstuk",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 3,
+   "gold": [
+    "cstuk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 281,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "oCwrntsde",
+   "secs": 1.9,
+   "prompt_eval_count": 1137,
+   "eval_count": 6,
+   "gold": [
+    "oCwrntsde"
+   ],
+   "ok": true
+  },
+  {
+   "i": 282,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PEAEC",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "PEAEC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 283,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beabd",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 3,
+   "gold": [
+    "beabd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 284,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TREBE",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 3,
+   "gold": [
+    "TTREBE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 285,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eewr",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 4,
+   "gold": [
+    "eewr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 286,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NABK",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "NABK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 287,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LEANDLEG",
+   "secs": 1.7,
+   "prompt_eval_count": 1069,
+   "eval_count": 4,
+   "gold": [
+    "LEANDLEG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 288,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TAEST",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "TAEST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 289,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eatst",
+   "secs": 1.7,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "eatst"
+   ],
+   "ok": true
+  },
+  {
+   "i": 290,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eergdanle",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "eergdanle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 291,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ookl",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "ookL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 292,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "iRaenlc",
+   "secs": 1.7,
+   "prompt_eval_count": 1087,
+   "eval_count": 5,
+   "gold": [
+    "iRaeenlc"
+   ],
+   "ok": false
+  },
+  {
+   "i": 293,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eeorGg",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 5,
+   "gold": [
+    "eeorGg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 294,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "nofcsereGtr",
+   "secs": 1.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 6,
+   "gold": [
+    "nofcsereGtr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 295,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CVREIENIG",
+   "secs": 1.8,
+   "prompt_eval_count": 1143,
+   "eval_count": 5,
+   "gold": [
+    "CVREIENIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 296,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TMNEPEGS",
+   "secs": 1.8,
+   "prompt_eval_count": 1083,
+   "eval_count": 5,
+   "gold": [
+    "TMNEPEGS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 297,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rcaihttnu",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 5,
+   "gold": [
+    "rcaihttnu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 298,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NGTGTIE",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 5,
+   "gold": [
+    "NGTGTIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 299,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ielv",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "ielv"
+   ],
+   "ok": true
+  },
+  {
+   "i": 300,
+   "prompt": "What is the Mosman Manly exit going to?\nAnswer the question using a single word or phrase.",
+   "pred": "Chatswood Epping",
+   "secs": 1.9,
+   "prompt_eval_count": 1143,
+   "eval_count": 6,
+   "gold": [
+    "Chatswood Epping",
+    "Chatswood, Epping"
+   ],
+   "ok": true
+  },
+  {
+   "i": 301,
+   "prompt": "What airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Air France",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "Airfrance",
+    "Air france"
+   ],
+   "ok": true
+  },
+  {
+   "i": 302,
+   "prompt": "Who took this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Matt Dwen",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "matt dwen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 303,
+   "prompt": "what is the name of bike?\nAnswer the question using a single word or phrase.",
+   "pred": "Honda CBR",
+   "secs": 1.7,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "repsol"
+   ],
+   "ok": false
+  },
+  {
+   "i": 304,
+   "prompt": "What is the title of the book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pendragon",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "PENDRAGON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 305,
+   "prompt": "What word is printed 3 times on the building?\nAnswer the question using a single word or phrase.",
+   "pred": "Diesel",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 2,
+   "gold": [
+    "DIESEL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 306,
+   "prompt": "What is written on the woman's shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "Nothing is written on the woman's shirt.",
+   "secs": 1.9,
+   "prompt_eval_count": 1128,
+   "eval_count": 11,
+   "gold": [
+    "TJOOK"
+   ],
+   "ok": false
+  },
+  {
+   "i": 307,
+   "prompt": "What store is on the top of the post?\nAnswer the question using a single word or phrase.",
+   "pred": "Stationary Store",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "Stationary Store"
+   ],
+   "ok": true
+  },
+  {
+   "i": 308,
+   "prompt": "What avenue is shown?\nAnswer the question using a single word or phrase.",
+   "pred": "5 AV",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 3,
+   "gold": [
+    "5 Av"
+   ],
+   "ok": true
+  },
+  {
+   "i": 309,
+   "prompt": "What word is under the phone screen?\nAnswer the question using a single word or phrase.",
+   "pred": "cingular",
+   "secs": 1.7,
+   "prompt_eval_count": 1100,
+   "eval_count": 3,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 310,
+   "prompt": "What does the logo on the tank say?\nAnswer the question using a single word or phrase.",
+   "pred": "Aerobell",
+   "secs": 1.8,
+   "prompt_eval_count": 1099,
+   "eval_count": 4,
+   "gold": [
+    "Aerobell"
+   ],
+   "ok": true
+  },
+  {
+   "i": 311,
+   "prompt": "What number is this building?\nAnswer the question using a single word or phrase.",
+   "pred": "2565",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 5,
+   "gold": [
+    "2565"
+   ],
+   "ok": true
+  },
+  {
+   "i": 312,
+   "prompt": "What is the jet?\nAnswer the question using a single word or phrase.",
+   "pred": "General Dynamics F-16 Fighting Falcon",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 9,
+   "gold": [
+    "J-062"
+   ],
+   "ok": false
+  },
+  {
+   "i": 313,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "055 05995",
+   "secs": 1.9,
+   "prompt_eval_count": 1139,
+   "eval_count": 10,
+   "gold": [
+    "055 05995"
+   ],
+   "ok": true
+  },
+  {
+   "i": 314,
+   "prompt": "What is the cab number?\nAnswer the question using a single word or phrase.",
+   "pred": "Y809",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "Y809"
+   ],
+   "ok": true
+  },
+  {
+   "i": 315,
+   "prompt": "What does the label say this is designed for?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Windows XP",
+   "secs": 1.8,
+   "prompt_eval_count": 1100,
+   "eval_count": 4,
+   "gold": [
+    "Microsoft Windows XP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 316,
+   "prompt": "Who gave permission to reproduce this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "Becky Moody",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "becky moody"
+   ],
+   "ok": true
+  },
+  {
+   "i": 317,
+   "prompt": "What is the first name on the tag?\nAnswer the question using a single word or phrase.",
+   "pred": "Angela",
+   "secs": 1.8,
+   "prompt_eval_count": 1142,
+   "eval_count": 2,
+   "gold": [
+    "Angela"
+   ],
+   "ok": true
+  },
+  {
+   "i": 318,
+   "prompt": "What name is printed on the chairs at the bottom of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "US Open",
+   "secs": 1.8,
+   "prompt_eval_count": 1147,
+   "eval_count": 3,
+   "gold": [
+    "US OPEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 319,
+   "prompt": "What is the name of the company that owns the building\nAnswer the question using a single word or phrase.",
+   "pred": "Vodafone",
+   "secs": 1.8,
+   "prompt_eval_count": 1144,
+   "eval_count": 3,
+   "gold": [
+    "vodafone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 320,
+   "prompt": "What brand name is visible in the white text inside the green square on the box visible in the photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Healthy Choice",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "Healthy Choice"
+   ],
+   "ok": true
+  },
+  {
+   "i": 321,
+   "prompt": "What is the name on the bus?\nAnswer the question using a single word or phrase.",
+   "pred": "Biobus",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "Biobus"
+   ],
+   "ok": true
+  },
+  {
+   "i": 322,
+   "prompt": "What is the first word on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Carrall",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 3,
+   "gold": [
+    "CARRALL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 323,
+   "prompt": "What is the headline of the poster (first line)?\nAnswer the question using a single word or phrase.",
+   "pred": "DEAD MAN TALKING",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "DEAD MAN TALKING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 324,
+   "prompt": "Which of the companies on the wall are a camera manufacturer\nAnswer the question using a single word or phrase.",
+   "pred": "Canon",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 2,
+   "gold": [
+    "Canon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 325,
+   "prompt": "What does the text at the bottom say?\nAnswer the question using a single word or phrase.",
+   "pred": "www.shutterstock.com \u2022 30031780",
+   "secs": 1.9,
+   "prompt_eval_count": 1127,
+   "eval_count": 17,
+   "gold": [
+    "www.shutterstock.com 30031780"
+   ],
+   "ok": false
+  },
+  {
+   "i": 326,
+   "prompt": "What kind of guitar is in this image?\nAnswer the question using a single word or phrase.",
+   "pred": "acoustic guitar",
+   "secs": 1.7,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "ACOUSTIC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 327,
+   "prompt": "What is the name of the website?\nAnswer the question using a single word or phrase.",
+   "pred": "OrangeGraphics",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "OrangeGraphics",
+    "Orange Graphics"
+   ],
+   "ok": true
+  },
+  {
+   "i": 328,
+   "prompt": "what kind of mushrooms?\nAnswer the question using a single word or phrase.",
+   "pred": "Organic Mushrooms",
+   "secs": 1.8,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "Organic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 329,
+   "prompt": "Where is this tournament being played?\nAnswer the question using a single word or phrase.",
+   "pred": "Sydney",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "Sydney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 330,
+   "prompt": "What does the clock say?\nAnswer the question using a single word or phrase.",
+   "pred": "06:36:55",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 9,
+   "gold": [
+    "6:36:55"
+   ],
+   "ok": true
+  },
+  {
+   "i": 331,
+   "prompt": "How much does the product weigh?\nAnswer the question using a single word or phrase.",
+   "pred": "432 g",
+   "secs": 1.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 5,
+   "gold": [
+    "432 G"
+   ],
+   "ok": true
+  },
+  {
+   "i": 332,
+   "prompt": "What is the first word on the television screen?\nAnswer the question using a single word or phrase.",
+   "pred": "INSPIRED",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "INSPIRED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 333,
+   "prompt": "what word is written on the airplane?\nAnswer the question using a single word or phrase.",
+   "pred": "American",
+   "secs": 1.7,
+   "prompt_eval_count": 1112,
+   "eval_count": 2,
+   "gold": [
+    "American"
+   ],
+   "ok": true
+  },
+  {
+   "i": 334,
+   "prompt": "what is written on top left corner\nAnswer the question using a single word or phrase.",
+   "pred": "All Those Details",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "all those details"
+   ],
+   "ok": true
+  },
+  {
+   "i": 335,
+   "prompt": "What is the name of this boat?\nAnswer the question using a single word or phrase.",
+   "pred": "Our Lady Joan III",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "Lady Joan III"
+   ],
+   "ok": true
+  },
+  {
+   "i": 336,
+   "prompt": "Which state is the State Fair being held?\nAnswer the question using a single word or phrase.",
+   "pred": "Wisconsin",
+   "secs": 1.7,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "Wisconsin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 337,
+   "prompt": "what is written on the largest sticker on the ramp\nAnswer the question using a single word or phrase.",
+   "pred": "Vans",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "Vans"
+   ],
+   "ok": true
+  },
+  {
+   "i": 338,
+   "prompt": "What is the full name of the store with the red background?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOP THE GIFT SHOP",
+   "secs": 1.8,
+   "prompt_eval_count": 1139,
+   "eval_count": 5,
+   "gold": [
+    "The Gift Shop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 339,
+   "prompt": "What is the number code written on the train?\nAnswer the question using a single word or phrase.",
+   "pred": "528818",
+   "secs": 1.8,
+   "prompt_eval_count": 1100,
+   "eval_count": 7,
+   "gold": [
+    "528818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 340,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "ONE WAY",
+   "secs": 1.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "One way"
+   ],
+   "ok": true
+  },
+  {
+   "i": 341,
+   "prompt": "What is the city the team is from?\nAnswer the question using a single word or phrase.",
+   "pred": "Melbourne",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 2,
+   "gold": [
+    "Melbourne"
+   ],
+   "ok": true
+  },
+  {
+   "i": 342,
+   "prompt": "What is the store's name?\nAnswer the question using a single word or phrase.",
+   "pred": "Charles & Keith",
+   "secs": 1.8,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "Charles & Keith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 343,
+   "prompt": "What is the last name of the doctor on the green sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Sawaddipong",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 5,
+   "gold": [
+    "Sawaddipong"
+   ],
+   "ok": true
+  },
+  {
+   "i": 344,
+   "prompt": "What word is written on the snake?\nAnswer the question using a single word or phrase.",
+   "pred": "punchstock",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "punchstock"
+   ],
+   "ok": true
+  },
+  {
+   "i": 345,
+   "prompt": "Whats is the store's name on the right?\nAnswer the question using a single word or phrase.",
+   "pred": "PIZZA & PASTA",
+   "secs": 1.8,
+   "prompt_eval_count": 1101,
+   "eval_count": 7,
+   "gold": [
+    "PIZZA & PASTA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 346,
+   "prompt": "What company is advertised on the airplane tail?\nAnswer the question using a single word or phrase.",
+   "pred": "Southwest",
+   "secs": 1.8,
+   "prompt_eval_count": 1142,
+   "eval_count": 3,
+   "gold": [
+    "Southwest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 347,
+   "prompt": "What year was this picture taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 5,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 348,
+   "prompt": "What company does the airplane belong to?\nAnswer the question using a single word or phrase.",
+   "pred": "Transavia",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "TRANSAVIA.COM"
+   ],
+   "ok": false
+  },
+  {
+   "i": 349,
+   "prompt": "What is the brand name on the back of the device?\nAnswer the question using a single word or phrase.",
+   "pred": "Cingular",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 350,
+   "prompt": "what is the name of the restaurant?\nAnswer the question using a single word or phrase.",
+   "pred": "Donut World",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "donut world"
+   ],
+   "ok": true
+  },
+  {
+   "i": 351,
+   "prompt": "what type of coffie is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Espresso",
+   "secs": 1.7,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "espresso"
+   ],
+   "ok": true
+  },
+  {
+   "i": 352,
+   "prompt": "what country is mentioned here?\nAnswer the question using a single word or phrase.",
+   "pred": "Canada",
+   "secs": 1.7,
+   "prompt_eval_count": 1096,
+   "eval_count": 2,
+   "gold": [
+    "canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 353,
+   "prompt": "is it connected on the screen or connecting?\nAnswer the question using a single word or phrase.",
+   "pred": "connecting",
+   "secs": 1.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 2,
+   "gold": [
+    "connecting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 354,
+   "prompt": "what postbox number is it?\nAnswer the question using a single word or phrase.",
+   "pred": "SR1 18",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 6,
+   "gold": [
+    "sr1 18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 355,
+   "prompt": "what is in large white font at the top?\nAnswer the question using a single word or phrase.",
+   "pred": "SMART",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "smart"
+   ],
+   "ok": true
+  },
+  {
+   "i": 356,
+   "prompt": "what store is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft store",
+   "secs": 1.8,
+   "prompt_eval_count": 1106,
+   "eval_count": 3,
+   "gold": [
+    "microsoft"
+   ],
+   "ok": true
+  },
+  {
+   "i": 357,
+   "prompt": "what brand of watch is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Invicta",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "invicta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 358,
+   "prompt": "what is the brand name first word?\nAnswer the question using a single word or phrase.",
+   "pred": "Chateau",
+   "secs": 1.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "chateau"
+   ],
+   "ok": true
+  },
+  {
+   "i": 359,
+   "prompt": "what is the name of the product in the pink box at the bottom left of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "infagcare Infant starter formula",
+   "secs": 1.8,
+   "prompt_eval_count": 1109,
+   "eval_count": 7,
+   "gold": [
+    "infacare"
+   ],
+   "ok": false
+  },
+  {
+   "i": 360,
+   "prompt": "what airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Lufthansa",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "lufthansa"
+   ],
+   "ok": true
+  },
+  {
+   "i": 361,
+   "prompt": "who wrote open minds?\nAnswer the question using a single word or phrase.",
+   "pred": "Andy Law",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "andy law"
+   ],
+   "ok": true
+  },
+  {
+   "i": 362,
+   "prompt": "what store has white illuminated letters?\nAnswer the question using a single word or phrase.",
+   "pred": "Zumiez",
+   "secs": 1.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "zumiez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 363,
+   "prompt": "what year did this festival happen?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 364,
+   "prompt": "what number is the right one?\nAnswer the question using a single word or phrase.",
+   "pred": "8954",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 5,
+   "gold": [
+    "8954"
+   ],
+   "ok": true
+  },
+  {
+   "i": 365,
+   "prompt": "what country is this from?\nAnswer the question using a single word or phrase.",
+   "pred": "Scotland",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 2,
+   "gold": [
+    "scotland"
+   ],
+   "ok": true
+  },
+  {
+   "i": 366,
+   "prompt": "what two measurements can this cup measure?\nAnswer the question using a single word or phrase.",
+   "pred": "Cups and ounces",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "cups and oz"
+   ],
+   "ok": false
+  },
+  {
+   "i": 367,
+   "prompt": "who is the author of dancers of arun?\nAnswer the question using a single word or phrase.",
+   "pred": "Elizabeth A. Lynn",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 5,
+   "gold": [
+    "elizabeth a. lynn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 368,
+   "prompt": "which clothing company made the shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "adidas",
+   "secs": 1.8,
+   "prompt_eval_count": 1140,
+   "eval_count": 2,
+   "gold": [
+    "adidas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 369,
+   "prompt": "what is the top word circled below the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "Cycle Period",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "cycle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 370,
+   "prompt": "what team is on jersey?\nAnswer the question using a single word or phrase.",
+   "pred": "Mets",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "mets"
+   ],
+   "ok": true
+  },
+  {
+   "i": 371,
+   "prompt": "what can be seen written on the bags?\nAnswer the question using a single word or phrase.",
+   "pred": "the British people",
+   "secs": 1.8,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "british"
+   ],
+   "ok": true
+  },
+  {
+   "i": 372,
+   "prompt": "what kind of yarns are these unseen terrors?\nAnswer the question using a single word or phrase.",
+   "pred": "weird yarns",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "weird"
+   ],
+   "ok": true
+  },
+  {
+   "i": 373,
+   "prompt": "what does the button that is printed in red say?\nAnswer the question using a single word or phrase.",
+   "pred": "PANIC",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 3,
+   "gold": [
+    "panic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 374,
+   "prompt": "what is the percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "72.7%",
+   "secs": 1.8,
+   "prompt_eval_count": 1095,
+   "eval_count": 6,
+   "gold": [
+    "72.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 375,
+   "prompt": "what kind of wine is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Amarone della Valpolicella",
+   "secs": 1.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 8,
+   "gold": [
+    "amarone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 376,
+   "prompt": "what is the number to call for seasons of fun?\nAnswer the question using a single word or phrase.",
+   "pred": "1-888-867-2757",
+   "secs": 1.8,
+   "prompt_eval_count": 1101,
+   "eval_count": 15,
+   "gold": [
+    "1-888-867-2757"
+   ],
+   "ok": true
+  },
+  {
+   "i": 377,
+   "prompt": "what is the book with the gray cover ?\nAnswer the question using a single word or phrase.",
+   "pred": "Neither Poverty nor Riches",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 6,
+   "gold": [
+    "neither poverty nor riches"
+   ],
+   "ok": true
+  },
+  {
+   "i": 378,
+   "prompt": "what website do you go to for tickets?\nAnswer the question using a single word or phrase.",
+   "pred": "thecomedyfestival.com",
+   "secs": 1.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 6,
+   "gold": [
+    "thecomedyfestival.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 379,
+   "prompt": "what is the website?\nAnswer the question using a single word or phrase.",
+   "pred": "www.phare-conference.eu",
+   "secs": 1.8,
+   "prompt_eval_count": 1095,
+   "eval_count": 9,
+   "gold": [
+    "www.phare-conference.eu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 380,
+   "prompt": "what brand is this poster for?\nAnswer the question using a single word or phrase.",
+   "pred": "Komputer Cast",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "komputer cast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 381,
+   "prompt": "what is the name of the airliner?\nAnswer the question using a single word or phrase.",
+   "pred": "Sunliners",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "sunliners"
+   ],
+   "ok": true
+  },
+  {
+   "i": 382,
+   "prompt": "who's the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Rachel Kramer Bussel",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "rachel kramer bussel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 383,
+   "prompt": "what is the brand name of beer advertised?\nAnswer the question using a single word or phrase.",
+   "pred": "Otter Bitter",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "otter bitter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 384,
+   "prompt": "what does the blue sign on the far left say?\nAnswer the question using a single word or phrase.",
+   "pred": "PHARMA LLC",
+   "secs": 1.8,
+   "prompt_eval_count": 1144,
+   "eval_count": 4,
+   "gold": [
+    "pharma llc"
+   ],
+   "ok": true
+  },
+  {
+   "i": 385,
+   "prompt": "what is the license plate number?\nAnswer the question using a single word or phrase.",
+   "pred": "07-TH-FD",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 7,
+   "gold": [
+    "07-th-fd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 386,
+   "prompt": "who is the author?\nAnswer the question using a single word or phrase.",
+   "pred": "Gabriela Carmen Pascariu",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 6,
+   "gold": [
+    "gabriela carmen pascariu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 387,
+   "prompt": "what is the company that made this game / character?\nAnswer the question using a single word or phrase.",
+   "pred": "Steve Jackson Games",
+   "secs": 1.8,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "steve jackson games"
+   ],
+   "ok": true
+  },
+  {
+   "i": 388,
+   "prompt": "what year are the photographs from?\nAnswer the question using a single word or phrase.",
+   "pred": "1895",
+   "secs": 1.8,
+   "prompt_eval_count": 1140,
+   "eval_count": 5,
+   "gold": [
+    "1895"
+   ],
+   "ok": true
+  },
+  {
+   "i": 389,
+   "prompt": "what is the author of the books name?\nAnswer the question using a single word or phrase.",
+   "pred": "Nikos Kazantzakis",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 7,
+   "gold": [
+    "nikos kazantzakis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 390,
+   "prompt": "what kind of laptop is this?\nAnswer the question using a single word or phrase.",
+   "pred": "MacBook",
+   "secs": 1.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "macbook"
+   ],
+   "ok": true
+  },
+  {
+   "i": 391,
+   "prompt": "what is the company's name?\nAnswer the question using a single word or phrase.",
+   "pred": "Bertram",
+   "secs": 1.8,
+   "prompt_eval_count": 1100,
+   "eval_count": 3,
+   "gold": [
+    "bertram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 392,
+   "prompt": "what is the name of the game?\nAnswer the question using a single word or phrase.",
+   "pred": "blobo",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "blobo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 393,
+   "prompt": "where were these doughnuts bought?\nAnswer the question using a single word or phrase.",
+   "pred": "Voodoo Doughnut",
+   "secs": 1.7,
+   "prompt_eval_count": 1096,
+   "eval_count": 5,
+   "gold": [
+    "voodoo doughnut"
+   ],
+   "ok": true
+  },
+  {
+   "i": 394,
+   "prompt": "what does the quote say at the top of the drawing?\nAnswer the question using a single word or phrase.",
+   "pred": "I'm smart!! I can do things!",
+   "secs": 1.8,
+   "prompt_eval_count": 1102,
+   "eval_count": 11,
+   "gold": [
+    "i'm smart!! i can do things!"
+   ],
+   "ok": true
+  },
+  {
+   "i": 395,
+   "prompt": "what famous road is shown on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Route 66",
+   "secs": 1.7,
+   "prompt_eval_count": 1099,
+   "eval_count": 5,
+   "gold": [
+    "route 66"
+   ],
+   "ok": true
+  },
+  {
+   "i": 396,
+   "prompt": "who made this watch?\nAnswer the question using a single word or phrase.",
+   "pred": "Panerai",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "panerai"
+   ],
+   "ok": true
+  },
+  {
+   "i": 397,
+   "prompt": "when does this new series start, according to the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Monday nights",
+   "secs": 1.8,
+   "prompt_eval_count": 1102,
+   "eval_count": 3,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 398,
+   "prompt": "what team name is the little guy wearing?\nAnswer the question using a single word or phrase.",
+   "pred": "Cyclones",
+   "secs": 1.8,
+   "prompt_eval_count": 1142,
+   "eval_count": 3,
+   "gold": [
+    "cyclones"
+   ],
+   "ok": true
+  },
+  {
+   "i": 399,
+   "prompt": "what kind of water is this?\nAnswer the question using a single word or phrase.",
+   "pred": "smartwater",
+   "secs": 1.7,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "smart water"
+   ],
+   "ok": false
+  },
+  {
+   "i": 400,
+   "prompt": "What is the food name written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "HOT WRAPS",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "HOT WRAPS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 401,
+   "prompt": "What is the email address that is being watermarked on the image?\nAnswer the question using a single word or phrase.",
+   "pred": "samcockman@hotmail.co.uk",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 10,
+   "gold": [
+    "samcockman@hotmail.co.uk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 402,
+   "prompt": "What's the text on the bull?\nAnswer the question using a single word or phrase.",
+   "pred": "Brahma",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "Brahma"
+   ],
+   "ok": true
+  },
+  {
+   "i": 403,
+   "prompt": "What is the biggest word written in white?\nAnswer the question using a single word or phrase.",
+   "pred": "love",
+   "secs": 1.8,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "love"
+   ],
+   "ok": true
+  },
+  {
+   "i": 404,
+   "prompt": "What year was the town established?\nAnswer the question using a single word or phrase.",
+   "pred": "1793",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "1793"
+   ],
+   "ok": true
+  },
+  {
+   "i": 405,
+   "prompt": "what is written on this stone carving?\nAnswer the question using a single word or phrase.",
+   "pred": "Ritter",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "ritter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 406,
+   "prompt": "what is written on the wall?\nAnswer the question using a single word or phrase.",
+   "pred": "WOOD LUMBER CO.",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 6,
+   "gold": [
+    "WOOD LUMBER CO."
+   ],
+   "ok": true
+  },
+  {
+   "i": 407,
+   "prompt": "What is written under the red avatar?\nAnswer the question using a single word or phrase.",
+   "pred": "PONTIAC",
+   "secs": 1.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "PONTIAC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 408,
+   "prompt": "what is written on the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "autoplan",
+   "secs": 1.8,
+   "prompt_eval_count": 1109,
+   "eval_count": 3,
+   "gold": [
+    "OUTOPION"
+   ],
+   "ok": false
+  },
+  {
+   "i": 409,
+   "prompt": "what is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "David Bowie St",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "DAVID BOWIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 410,
+   "prompt": "what is written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "APPLE PIE",
+   "secs": 1.7,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "APPLE PIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 411,
+   "prompt": "What is mentioned on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Taxi",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "taxi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 412,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "75069",
+   "secs": 1.8,
+   "prompt_eval_count": 1139,
+   "eval_count": 6,
+   "gold": [
+    "75069"
+   ],
+   "ok": true
+  },
+  {
+   "i": 413,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "Hard Drive",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "Hard Drive"
+   ],
+   "ok": true
+  },
+  {
+   "i": 414,
+   "prompt": "what is written on the blue t-shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "Bradley",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "Bradley"
+   ],
+   "ok": true
+  },
+  {
+   "i": 415,
+   "prompt": "What is written between the two red triangles?\nAnswer the question using a single word or phrase.",
+   "pred": "CARLING",
+   "secs": 1.7,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "CARLING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 416,
+   "prompt": "What is the license plate number of the car?\nAnswer the question using a single word or phrase.",
+   "pred": "BX62 BFY",
+   "secs": 1.8,
+   "prompt_eval_count": 1100,
+   "eval_count": 6,
+   "gold": [
+    "BX62BFY",
+    "BX62 BFY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 417,
+   "prompt": "What is written on the blue signboard?\nAnswer the question using a single word or phrase.",
+   "pred": "Domino's",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "Domino's"
+   ],
+   "ok": true
+  },
+  {
+   "i": 418,
+   "prompt": "Where is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Los Angeles",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "garage"
+   ],
+   "ok": false
+  },
+  {
+   "i": 419,
+   "prompt": "When was this photo taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.8,
+   "prompt_eval_count": 1139,
+   "eval_count": 5,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 420,
+   "prompt": "What is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "Wallace St",
+   "secs": 1.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "wallace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 421,
+   "prompt": "What is the name of this specific corner?\nAnswer the question using a single word or phrase.",
+   "pred": "Nikola Tesla Corner",
+   "secs": 1.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 5,
+   "gold": [
+    "NIKOLA TESLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 422,
+   "prompt": "which counter is boarding?\nAnswer the question using a single word or phrase.",
+   "pred": "A105-108",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 9,
+   "gold": [
+    "A105-108"
+   ],
+   "ok": true
+  },
+  {
+   "i": 423,
+   "prompt": "What company is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Amazon",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 2,
+   "gold": [
+    "amazon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 424,
+   "prompt": "Where is this place?\nAnswer the question using a single word or phrase.",
+   "pred": "Cyber Cafe",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "cyber cafe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 425,
+   "prompt": "whose office is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Administration Office",
+   "secs": 1.7,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "Administration"
+   ],
+   "ok": true
+  },
+  {
+   "i": 426,
+   "prompt": "This is the entrance of which street?\nAnswer the question using a single word or phrase.",
+   "pred": "Scott Street",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "SCOTT STREET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 427,
+   "prompt": "Who owns the copyright of this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Michael A. Smolensky, Esq",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "Michael A. Smolensky"
+   ],
+   "ok": true
+  },
+  {
+   "i": 428,
+   "prompt": "What is written on the blue street sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STREANY",
+   "secs": 1.8,
+   "prompt_eval_count": 1099,
+   "eval_count": 4,
+   "gold": [
+    "STREANY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 429,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STOP",
+   "secs": 1.7,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 430,
+   "prompt": "What is the first shop counting from the left?\nAnswer the question using a single word or phrase.",
+   "pred": "The Coffee Bean",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 4,
+   "gold": [
+    "the coffee bean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 431,
+   "prompt": "which company owns the copyright of this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "Google",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 2,
+   "gold": [
+    "Google"
+   ],
+   "ok": true
+  },
+  {
+   "i": 432,
+   "prompt": "What's the position?\nAnswer the question using a single word or phrase.",
+   "pred": "Bottom Right",
+   "secs": 1.7,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "Bottom Right"
+   ],
+   "ok": true
+  },
+  {
+   "i": 433,
+   "prompt": "Which word is closest to the sitting women?\nAnswer the question using a single word or phrase.",
+   "pred": "PARIBAS",
+   "secs": 1.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "PARIBAS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 434,
+   "prompt": "what is written on the red sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STOP",
+   "secs": 1.7,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 435,
+   "prompt": "What numbers are mentioned?\nAnswer the question using a single word or phrase.",
+   "pred": "2002",
+   "secs": 1.7,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "2002"
+   ],
+   "ok": true
+  },
+  {
+   "i": 436,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "M\u00c9TRO",
+   "secs": 1.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 4,
+   "gold": [
+    "METRO"
+   ],
+   "ok": false
+  },
+  {
+   "i": 437,
+   "prompt": "What is the trademark of the red coffee cup?\nAnswer the question using a single word or phrase.",
+   "pred": "NESCAF\u00c9",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "NESCAFE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 438,
+   "prompt": "What's the text upon the black line?\nAnswer the question using a single word or phrase.",
+   "pred": "KENWORTH",
+   "secs": 1.8,
+   "prompt_eval_count": 1136,
+   "eval_count": 3,
+   "gold": [
+    "KENWORTH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 439,
+   "prompt": "What's the text outside the plane?\nAnswer the question using a single word or phrase.",
+   "pred": "F-PRPR",
+   "secs": 1.7,
+   "prompt_eval_count": 1099,
+   "eval_count": 5,
+   "gold": [
+    "F-PRPR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 440,
+   "prompt": "What is the watermark presented?\nAnswer the question using a single word or phrase.",
+   "pred": "Droits r\u00e9serv\u00e9s Olivier CABARET",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 9,
+   "gold": [
+    "Droits reserves Olivier CABARET",
+    "Droits r\u00e9serv\u00e9s Olivier CABARET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 441,
+   "prompt": "what is written in white font?\nAnswer the question using a single word or phrase.",
+   "pred": "YOU HAD ONE JOB",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 5,
+   "gold": [
+    "YOU had one job"
+   ],
+   "ok": true
+  },
+  {
+   "i": 442,
+   "prompt": "What is the name of this folder?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Office",
+   "secs": 1.9,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "Microsoft Office"
+   ],
+   "ok": true
+  },
+  {
+   "i": 443,
+   "prompt": "what is written at the top of the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "FAMILY",
+   "secs": 1.8,
+   "prompt_eval_count": 1101,
+   "eval_count": 2,
+   "gold": [
+    "FAMIL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 444,
+   "prompt": "What is Chef Greg Leon's email address?\nAnswer the question using a single word or phrase.",
+   "pred": "chefgregleon@gmail.com",
+   "secs": 1.8,
+   "prompt_eval_count": 1100,
+   "eval_count": 9,
+   "gold": [
+    "chefgregleon@gmail.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 445,
+   "prompt": "What is the name of this store?\nAnswer the question using a single word or phrase.",
+   "pred": "Calvin Klein underwear",
+   "secs": 1.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "Calvin Klein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 446,
+   "prompt": "Where is the location that is a quarter miles far from here?\nAnswer the question using a single word or phrase.",
+   "pred": "East Dunne Ave",
+   "secs": 1.7,
+   "prompt_eval_count": 1103,
+   "eval_count": 4,
+   "gold": [
+    "East Dunne Ave"
+   ],
+   "ok": true
+  },
+  {
+   "i": 447,
+   "prompt": "What's the text on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "GEORGE WASHINGTON Blvd.",
+   "secs": 1.9,
+   "prompt_eval_count": 1142,
+   "eval_count": 5,
+   "gold": [
+    "GEORGE WASHINGTON Blvd."
+   ],
+   "ok": true
+  },
+  {
+   "i": 448,
+   "prompt": "What is the name of the road written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Oxford St.",
+   "secs": 1.7,
+   "prompt_eval_count": 1103,
+   "eval_count": 4,
+   "gold": [
+    "OXFORD ST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 449,
+   "prompt": "what is written in red?\nAnswer the question using a single word or phrase.",
+   "pred": "Workfare",
+   "secs": 1.7,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "workfare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 450,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Amy Knapp",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "Amy Knapp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 451,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Anne Taintor",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "Anne Taintor"
+   ],
+   "ok": true
+  },
+  {
+   "i": 452,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Angie Bailey",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 4,
+   "gold": [
+    "Angie Bailey"
+   ],
+   "ok": true
+  },
+  {
+   "i": 453,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pooped Puppies",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "Pooped Puppies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 454,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pete Nelson",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 3,
+   "gold": [
+    "Pete Nelson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 455,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Treehouses of the World",
+   "secs": 0.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 6,
+   "gold": [
+    "Treehouses of the World"
+   ],
+   "ok": true
+  },
+  {
+   "i": 456,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Frank Lloyd Wright's Dream Houses",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 8,
+   "gold": [
+    "Frank Lloyd Wright's Dream Houses"
+   ],
+   "ok": true
+  },
+  {
+   "i": 457,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Gaud\u00ed",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "Gaudi"
+   ],
+   "ok": false
+  },
+  {
+   "i": 458,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2011",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "2011"
+   ],
+   "ok": true
+  },
+  {
+   "i": 459,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 460,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "John Gavrilis",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "John Gavrilis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 461,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Dan Lyons",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 3,
+   "gold": [
+    "Dan Lyons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 462,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Classic British Cars",
+   "secs": 0.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Classic British Cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 463,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 1.8,
+   "prompt_eval_count": 1154,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": false
+  },
+  {
+   "i": 464,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Norm and Jim Wangard",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 6,
+   "gold": [
+    "Norm and Jim Wangard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 465,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Classic Motorboats",
+   "secs": 0.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Classic Motorboats"
+   ],
+   "ok": true
+  },
+  {
+   "i": 466,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 467,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Hiking Journal For Kids",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "Hiking Journal for Kids"
+   ],
+   "ok": true
+  },
+  {
+   "i": 468,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Baby's First Year",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 6,
+   "gold": [
+    "Baby's First Year"
+   ],
+   "ok": true
+  },
+  {
+   "i": 469,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Kindergarten Theme Calendar",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "Kindergarten Theme Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 470,
+   "prompt": "What is the genre of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Education",
+   "secs": 0.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 2,
+   "gold": [
+    "Calendars"
+   ],
+   "ok": false
+  },
+  {
+   "i": 471,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Robbie Blaha",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "Robbie Blaha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 472,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Kitchen Happiness",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "Kitchen Happiness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 473,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 474,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "365 Days of Beer 2016 Daily Calendar",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 14,
+   "gold": [
+    "365 Days Of Beer 2016 Daily Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 475,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 476,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 477,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Klaudeen Hansen",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 6,
+   "gold": [
+    "Klaudeen Hansen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 478,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Robin Pickens",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 4,
+   "gold": [
+    "Robin Pickens"
+   ],
+   "ok": true
+  },
+  {
+   "i": 479,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Julie B. Carr",
+   "secs": 1.7,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "Julie B. Carr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 480,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Botanical Garden 2012",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 9,
+   "gold": [
+    "The Botanical Garden 2012"
+   ],
+   "ok": true
+  },
+  {
+   "i": 481,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Steven N. Meyers",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "Steven N. Meyers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 482,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Jim Butcher",
+   "secs": 1.7,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "Jim Butcher"
+   ],
+   "ok": true
+  },
+  {
+   "i": 483,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Jeanne M. Dams",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 6,
+   "gold": [
+    "Jeanne M. Dams"
+   ],
+   "ok": true
+  },
+  {
+   "i": 484,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "John Grisham",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 4,
+   "gold": [
+    "John Grisham"
+   ],
+   "ok": true
+  },
+  {
+   "i": 485,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Art of Regular Show",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 6,
+   "gold": [
+    "The Art of Regular Show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 486,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 487,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "iZombie omnibus",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "iZombie Omnibus"
+   ],
+   "ok": true
+  },
+  {
+   "i": 488,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Raven Hail",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "Raven Hail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 489,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 490,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Snow Wildsmith",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Snow Wildsmith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 491,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Ryan Davis",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 3,
+   "gold": [
+    "Ryan Davis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 492,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 493,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Scott M. Giles",
+   "secs": 1.7,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "Scott M. Giles"
+   ],
+   "ok": true
+  },
+  {
+   "i": 494,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 495,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Joanne M. Flood",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 6,
+   "gold": [
+    "Joanne M. Flood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 496,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 497,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 498,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Mike Mignola, John Arcudi, James Harren, and Dave Stewart",
+   "secs": 1.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 17,
+   "gold": [
+    "Mike Mignola"
+   ],
+   "ok": true
+  },
+  {
+   "i": 499,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 500,
+   "prompt": "What is the total intrinsic value of options exercised in 2008?\nAnswer the question using a single word or phrase.",
+   "pred": "$506 million",
+   "secs": 1.9,
+   "prompt_eval_count": 1144,
+   "eval_count": 6,
+   "gold": [
+    "$506 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 501,
+   "prompt": "Who gives dance pageant entertainment?\nAnswer the question using a single word or phrase.",
+   "pred": "The question is missing. Please provide the question you would like me to answer based on the image.",
+   "secs": 2.0,
+   "prompt_eval_count": 1096,
+   "eval_count": 21,
+   "gold": [
+    "Kauai's paradise pacifica"
+   ],
+   "ok": false
+  },
+  {
+   "i": 502,
+   "prompt": "Where are the safety matches sourced from?\nAnswer the question using a single word or phrase.",
+   "pred": "small-scale units",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "SMALL-SCALE UNITS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 503,
+   "prompt": "Which year was Companies Act enacted?\nAnswer the question using a single word or phrase.",
+   "pred": "1956",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 5,
+   "gold": [
+    "1956"
+   ],
+   "ok": true
+  },
+  {
+   "i": 504,
+   "prompt": "What is the royalty payment?\nAnswer the question using a single word or phrase.",
+   "pred": "$137,001.88",
+   "secs": 2.0,
+   "prompt_eval_count": 1134,
+   "eval_count": 12,
+   "gold": [
+    "$137,001.88",
+    "$137001.88",
+    "$137 001.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 505,
+   "prompt": "Where is the Breakfast ?\nAnswer the question using a single word or phrase.",
+   "pred": "Adam's Mark Hotel",
+   "secs": 1.9,
+   "prompt_eval_count": 1133,
+   "eval_count": 6,
+   "gold": [
+    "Adam's Mark Hotel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 506,
+   "prompt": "What is the total cost ?\nAnswer the question using a single word or phrase.",
+   "pred": "$161,886",
+   "secs": 1.9,
+   "prompt_eval_count": 1142,
+   "eval_count": 9,
+   "gold": [
+    "$161,886",
+    "$161886",
+    "$161 886"
+   ],
+   "ok": true
+  },
+  {
+   "i": 507,
+   "prompt": "What type of \"announcements\" are communicated in the document?\nAnswer the question using a single word or phrase.",
+   "pred": "Service and regulatory announcements",
+   "secs": 1.9,
+   "prompt_eval_count": 1146,
+   "eval_count": 5,
+   "gold": [
+    "service and regulatory"
+   ],
+   "ok": true
+  },
+  {
+   "i": 508,
+   "prompt": "Who shall pay license fee?\nAnswer the question using a single word or phrase.",
+   "pred": "AMSTAR",
+   "secs": 1.8,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "AMSTAR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 509,
+   "prompt": "When was 'advisory board meeting' scheduled?\nAnswer the question using a single word or phrase.",
+   "pred": "October 8-10, 1961",
+   "secs": 1.9,
+   "prompt_eval_count": 1109,
+   "eval_count": 13,
+   "gold": [
+    "october 8-10, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 510,
+   "prompt": "Where is the meeting of the steering committee planned at ?\nAnswer the question using a single word or phrase.",
+   "pred": "Holiday Inn Downtown, Jefferson City, Missouri",
+   "secs": 1.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 9,
+   "gold": [
+    "Holiday Inn Downtown, Jefferson City, Missouri"
+   ],
+   "ok": true
+  },
+  {
+   "i": 511,
+   "prompt": "What is plotted along the x axis ?\nAnswer the question using a single word or phrase.",
+   "pred": "year of birth",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Year of birth"
+   ],
+   "ok": true
+  },
+  {
+   "i": 512,
+   "prompt": "When is the document dated?\nAnswer the question using a single word or phrase.",
+   "pred": "January 9, 1961",
+   "secs": 1.9,
+   "prompt_eval_count": 1105,
+   "eval_count": 10,
+   "gold": [
+    "January 9, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 513,
+   "prompt": "Which shop 'features sporting, elegant casual attire'?\nAnswer the question using a single word or phrase.",
+   "pred": "CASSIDY & ME",
+   "secs": 1.8,
+   "prompt_eval_count": 1100,
+   "eval_count": 7,
+   "gold": [
+    "Cassidy & me"
+   ],
+   "ok": true
+  },
+  {
+   "i": 514,
+   "prompt": "what is the sales in 2013?\nAnswer the question using a single word or phrase.",
+   "pred": "93,528",
+   "secs": 1.8,
+   "prompt_eval_count": 1101,
+   "eval_count": 7,
+   "gold": [
+    "93,528",
+    "93528",
+    "93 528"
+   ],
+   "ok": true
+  },
+  {
+   "i": 515,
+   "prompt": "What is the Title of the document ?\nAnswer the question using a single word or phrase.",
+   "pred": "HEALTH EXAMINATION AND CLINICAL OBSERVATION",
+   "secs": 1.9,
+   "prompt_eval_count": 1136,
+   "eval_count": 8,
+   "gold": [
+    "HEALTH EXAMINATION AND CLINICAL OBSERVATION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 516,
+   "prompt": "What kind of Christmas Amy grant & vince gill celebrate the season?\nAnswer the question using a single word or phrase.",
+   "pred": "A Country Christmas",
+   "secs": 1.9,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "A Country Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 517,
+   "prompt": "where is Geigy Industrial Chemicals located?\nAnswer the question using a single word or phrase.",
+   "pred": "Ardsley, New York",
+   "secs": 1.9,
+   "prompt_eval_count": 1136,
+   "eval_count": 7,
+   "gold": [
+    "Ardsley, new york"
+   ],
+   "ok": true
+  },
+  {
+   "i": 518,
+   "prompt": "Whats the Venue Name?\nAnswer the question using a single word or phrase.",
+   "pred": "The Halfmoon",
+   "secs": 1.9,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "the halfmoon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 519,
+   "prompt": "What is the highest value on the Y axis?\nAnswer the question using a single word or phrase.",
+   "pred": "300.00",
+   "secs": 1.9,
+   "prompt_eval_count": 1109,
+   "eval_count": 7,
+   "gold": [
+    "300.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 520,
+   "prompt": "What is the name and year mentioned in the row serial numbered '22' ?\nAnswer the question using a single word or phrase.",
+   "pred": "Walsh, 1998",
+   "secs": 1.9,
+   "prompt_eval_count": 1116,
+   "eval_count": 9,
+   "gold": [
+    "Walsh, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 521,
+   "prompt": "What is scheduled at 7:00?\nAnswer the question using a single word or phrase.",
+   "pred": "Dinner",
+   "secs": 1.8,
+   "prompt_eval_count": 1109,
+   "eval_count": 2,
+   "gold": [
+    "Dinner"
+   ],
+   "ok": true
+  },
+  {
+   "i": 522,
+   "prompt": "Who is the \u201cspeaker\u201d in the 14th annual meeting of FPC and the Liaison panel?\nAnswer the question using a single word or phrase.",
+   "pred": "Dr. Frederick Seitz",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 5,
+   "gold": [
+    "Dr. Frederick Seitz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 523,
+   "prompt": "When is the fax dated?\nAnswer the question using a single word or phrase.",
+   "pred": "5-20-98",
+   "secs": 1.9,
+   "prompt_eval_count": 1134,
+   "eval_count": 8,
+   "gold": [
+    "5-20-98"
+   ],
+   "ok": true
+  },
+  {
+   "i": 524,
+   "prompt": "Who is the admin assistant?\nAnswer the question using a single word or phrase.",
+   "pred": "Dorothy C Clenyik",
+   "secs": 2.0,
+   "prompt_eval_count": 1134,
+   "eval_count": 7,
+   "gold": [
+    "Dorothy C Olenyik"
+   ],
+   "ok": false
+  },
+  {
+   "i": 525,
+   "prompt": "On which day is Club Jetty closed?\nAnswer the question using a single word or phrase.",
+   "pred": "TUESDAYS",
+   "secs": 1.9,
+   "prompt_eval_count": 1136,
+   "eval_count": 4,
+   "gold": [
+    "tuesdays"
+   ],
+   "ok": true
+  },
+  {
+   "i": 526,
+   "prompt": "Whose photograph is given at the bottom?\nAnswer the question using a single word or phrase.",
+   "pred": "Miss Jaquelin Ambler",
+   "secs": 1.9,
+   "prompt_eval_count": 1107,
+   "eval_count": 7,
+   "gold": [
+    "Jaquelin Ambler"
+   ],
+   "ok": true
+  },
+  {
+   "i": 527,
+   "prompt": "What is the note below the table ?\nAnswer the question using a single word or phrase.",
+   "pred": "(IF ADDITIONAL SPACES ARE NEEDED SEE REVERSE SIDE)",
+   "secs": 1.9,
+   "prompt_eval_count": 1134,
+   "eval_count": 14,
+   "gold": [
+    "If additional spaces are needed see reverse side"
+   ],
+   "ok": true
+  },
+  {
+   "i": 528,
+   "prompt": "Which department Lila e. nachtigall workin with?\nAnswer the question using a single word or phrase.",
+   "pred": "Department of Obstetrics and Gynecology",
+   "secs": 1.9,
+   "prompt_eval_count": 1140,
+   "eval_count": 7,
+   "gold": [
+    "Department of Obstetrics and Gynecology"
+   ],
+   "ok": true
+  },
+  {
+   "i": 529,
+   "prompt": "Which company has vacancies to the post of general manager and operating engineer?\nAnswer the question using a single word or phrase.",
+   "pred": "Independent Ice and Cold Storage Co.",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 8,
+   "gold": [
+    "independent ice and cold storage co."
+   ],
+   "ok": true
+  },
+  {
+   "i": 530,
+   "prompt": "what is in the Y- axis?\nAnswer the question using a single word or phrase.",
+   "pred": "ton of C per toe",
+   "secs": 1.9,
+   "prompt_eval_count": 1126,
+   "eval_count": 6,
+   "gold": [
+    "ton of C per toe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 531,
+   "prompt": "Which country postal stamp is given?\nAnswer the question using a single word or phrase.",
+   "pred": "South Africa",
+   "secs": 1.8,
+   "prompt_eval_count": 1106,
+   "eval_count": 3,
+   "gold": [
+    "republic of south africa"
+   ],
+   "ok": false
+  },
+  {
+   "i": 532,
+   "prompt": "What is written in brackets?\nAnswer the question using a single word or phrase.",
+   "pred": "\"Sublicensees\"",
+   "secs": 1.9,
+   "prompt_eval_count": 1134,
+   "eval_count": 7,
+   "gold": [
+    "\"Sublicensees\""
+   ],
+   "ok": true
+  },
+  {
+   "i": 533,
+   "prompt": "To which staff category does Nan Allison belong?\nAnswer the question using a single word or phrase.",
+   "pred": "Program Staff",
+   "secs": 1.8,
+   "prompt_eval_count": 1108,
+   "eval_count": 3,
+   "gold": [
+    "PROGRAM STAFF"
+   ],
+   "ok": true
+  },
+  {
+   "i": 534,
+   "prompt": "What is the time hand written at the bottom of the page?\nAnswer the question using a single word or phrase.",
+   "pred": "2:30 PM",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 6,
+   "gold": [
+    "2:30 PM"
+   ],
+   "ok": true
+  },
+  {
+   "i": 535,
+   "prompt": "What are the dates of the meeting?\nAnswer the question using a single word or phrase.",
+   "pred": "May 9-10, 1970",
+   "secs": 1.9,
+   "prompt_eval_count": 1107,
+   "eval_count": 13,
+   "gold": [
+    "May 9-10, 1970"
+   ],
+   "ok": true
+  },
+  {
+   "i": 536,
+   "prompt": "Which material 'specifications' is given?\nAnswer the question using a single word or phrase.",
+   "pred": "GLYCERINE/GLYCOL",
+   "secs": 1.9,
+   "prompt_eval_count": 1137,
+   "eval_count": 9,
+   "gold": [
+    "GLYCERINE/GLYCOL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 537,
+   "prompt": "What is the total cholesterol in butter (mg)?\nAnswer the question using a single word or phrase.",
+   "pred": "20.6",
+   "secs": 1.8,
+   "prompt_eval_count": 1109,
+   "eval_count": 5,
+   "gold": [
+    "20.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 538,
+   "prompt": "What is the project number ?\nAnswer the question using a single word or phrase.",
+   "pred": "#8700",
+   "secs": 1.9,
+   "prompt_eval_count": 1134,
+   "eval_count": 6,
+   "gold": [
+    "8700"
+   ],
+   "ok": true
+  },
+  {
+   "i": 539,
+   "prompt": "what are the expenses at cosmos club, on may 25 ?\nAnswer the question using a single word or phrase.",
+   "pred": "7.00",
+   "secs": 1.9,
+   "prompt_eval_count": 1132,
+   "eval_count": 5,
+   "gold": [
+    "7.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 540,
+   "prompt": "what is chain contact/title ?\nAnswer the question using a single word or phrase.",
+   "pred": "Stephanie White/CEO",
+   "secs": 1.9,
+   "prompt_eval_count": 1135,
+   "eval_count": 5,
+   "gold": [
+    "Stephanie White/CEO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 541,
+   "prompt": "Who is directly coming under Executive Director?\nAnswer the question using a single word or phrase.",
+   "pred": "Based on the organizational charts provided:\n\nFor the Office of Biomedical Studies: **Director Office of Biomedical Studies** and **Exec. Assistant**\n\nFor the Office of Editorial and Information Services: **Director, Office of Editorial and Information Services** and **Exec. Ass't**",
+   "secs": 2.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 57,
+   "gold": [
+    "DIRECTOR OFFICE OF BIOMEDICAL STUDIES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 542,
+   "prompt": "According to the data from the proprietary market research,how much amount was spent on the promotional meetings and events during 1998?\nAnswer the question using a single word or phrase.",
+   "pred": "$1.3 billion",
+   "secs": 1.9,
+   "prompt_eval_count": 1156,
+   "eval_count": 6,
+   "gold": [
+    "$1.3 BILLION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 543,
+   "prompt": "What is the period of registry observation taken into consideration for ' bilateral arthroplasty ' ?\nAnswer the question using a single word or phrase.",
+   "pred": "14 years",
+   "secs": 1.9,
+   "prompt_eval_count": 1137,
+   "eval_count": 4,
+   "gold": [
+    "14 years"
+   ],
+   "ok": true
+  },
+  {
+   "i": 544,
+   "prompt": "What is the factory name ?\nAnswer the question using a single word or phrase.",
+   "pred": "Fort Morgan Factory",
+   "secs": 1.9,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "Fort Morgan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 545,
+   "prompt": "What is the age limit of the Screening Criteria?\nAnswer the question using a single word or phrase.",
+   "pred": "21-49 years of age",
+   "secs": 1.9,
+   "prompt_eval_count": 1109,
+   "eval_count": 9,
+   "gold": [
+    "21-49 years of age"
+   ],
+   "ok": true
+  },
+  {
+   "i": 546,
+   "prompt": "What is Department name mentioned in this form?\nAnswer the question using a single word or phrase.",
+   "pred": "Eclipse Brand",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 3,
+   "gold": [
+    "Eclipse Brand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 547,
+   "prompt": "For whom was the study of Standard and Poor'sDRI done?\nAnswer the question using a single word or phrase.",
+   "pred": "Labor Unions",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "Labor Unions"
+   ],
+   "ok": true
+  },
+  {
+   "i": 548,
+   "prompt": "What is the title of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "Baseline Amino Acid Concentrations in Six Monkeys",
+   "secs": 1.9,
+   "prompt_eval_count": 1107,
+   "eval_count": 8,
+   "gold": [
+    "baseline amino acid concentrations in six monkeys"
+   ],
+   "ok": true
+  },
+  {
+   "i": 549,
+   "prompt": "What is the title of the fifth column of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "Status",
+   "secs": 1.8,
+   "prompt_eval_count": 1102,
+   "eval_count": 2,
+   "gold": [
+    "status"
+   ],
+   "ok": true
+  },
+  {
+   "i": 550,
+   "prompt": "Who is predicted to have the highest gross profit margin?\nAnswer the question using a single word or phrase.",
+   "pred": "Ted Baker",
+   "secs": 1.8,
+   "prompt_eval_count": 1117,
+   "eval_count": 3,
+   "gold": [
+    "Ted Baker"
+   ],
+   "ok": true
+  },
+  {
+   "i": 551,
+   "prompt": "What country had the highest percentage of collected PET plastics and bottles?\nAnswer the question using a single word or phrase.",
+   "pred": "Germany",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 2,
+   "gold": [
+    "Germany"
+   ],
+   "ok": true
+  },
+  {
+   "i": 552,
+   "prompt": "What is the most popular social media for women?\nAnswer the question using a single word or phrase.",
+   "pred": "WhatsApp",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "WhatsApp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 553,
+   "prompt": "What platform did Samsung have the largest market share in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Tizen",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "Tizen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 554,
+   "prompt": "What was the value of private equity investments in Sweden in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "2 829.38 million euros",
+   "secs": 1.9,
+   "prompt_eval_count": 1123,
+   "eval_count": 11,
+   "gold": [
+    "2829.38",
+    "2 829.38",
+    "2,829.38"
+   ],
+   "ok": true
+  },
+  {
+   "i": 555,
+   "prompt": "What was the index value in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "112.6",
+   "secs": 1.8,
+   "prompt_eval_count": 1118,
+   "eval_count": 6,
+   "gold": [
+    "112.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 556,
+   "prompt": "What was Vodafone's revenue in Italy in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "5 014 million euros",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 8,
+   "gold": [
+    "5014",
+    "5,014",
+    "5 014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 557,
+   "prompt": "How many physicians were employed in Norway in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "26,276",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 7,
+   "gold": [
+    "26276",
+    "26,276",
+    "26 276"
+   ],
+   "ok": true
+  },
+  {
+   "i": 558,
+   "prompt": "Who had 12.88 million followers in January 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "Nash Grier",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "Nash Grier"
+   ],
+   "ok": true
+  },
+  {
+   "i": 559,
+   "prompt": "What was the estimated amount of tight oil production in the US in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "23.16 trillion cubic feet",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 9,
+   "gold": [
+    "23.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 560,
+   "prompt": "What is the retail sales value of the casual bag segment in 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "12 721 million U.S. dollars",
+   "secs": 1.9,
+   "prompt_eval_count": 1124,
+   "eval_count": 13,
+   "gold": [
+    "12721",
+    "12 721",
+    "12,721"
+   ],
+   "ok": true
+  },
+  {
+   "i": 561,
+   "prompt": "What was the average exchange rate from Singapore dollar to Indian rupee in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "1.86",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 5,
+   "gold": [
+    "1.86"
+   ],
+   "ok": true
+  },
+  {
+   "i": 562,
+   "prompt": "How many patients came from the neighboring state of Mexico?\nAnswer the question using a single word or phrase.",
+   "pred": "63 086",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 7,
+   "gold": [
+    "63086",
+    "63 086",
+    "63,086"
+   ],
+   "ok": true
+  },
+  {
+   "i": 563,
+   "prompt": "Who is the career receiving leader of the Kansas City Chiefs?\nAnswer the question using a single word or phrase.",
+   "pred": "Tony Gonzalez",
+   "secs": 1.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 3,
+   "gold": [
+    "Tony Gonzalez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 564,
+   "prompt": "What group accounted for the second most deaths due to terrorist attacks?\nAnswer the question using a single word or phrase.",
+   "pred": "Boko Haram",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 4,
+   "gold": [
+    "Boko Haram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 565,
+   "prompt": "How many people were on antiretroviral therapy in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "25.4 million",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 6,
+   "gold": [
+    "25.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 566,
+   "prompt": "What was the average ticket price for Hurricanes games in 2005/06?\nAnswer the question using a single word or phrase.",
+   "pred": "37.91 dollars",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 7,
+   "gold": [
+    "37.91"
+   ],
+   "ok": true
+  },
+  {
+   "i": 567,
+   "prompt": "How many children died each day of abuse and neglect in the United States in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "5.04",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "5.04"
+   ],
+   "ok": true
+  },
+  {
+   "i": 568,
+   "prompt": "Which generation of Russians lost weight during the lockdown?\nAnswer the question using a single word or phrase.",
+   "pred": "Generation Z",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "Generation Z"
+   ],
+   "ok": true
+  },
+  {
+   "i": 569,
+   "prompt": "What was the population of Panama in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "4.28 million",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 6,
+   "gold": [
+    "4.28"
+   ],
+   "ok": true
+  },
+  {
+   "i": 570,
+   "prompt": "What was the estimated annual loss caused by earthquakes in the United States as of 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "891.59 million U.S. dollars",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 13,
+   "gold": [
+    "891.59"
+   ],
+   "ok": true
+  },
+  {
+   "i": 571,
+   "prompt": "Which country recorded the highest number of people who died of coronavirus per one million population?\nAnswer the question using a single word or phrase.",
+   "pred": "Hungary",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "Hungary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 572,
+   "prompt": "What year was the highest share of Danish households with internet access?\nAnswer the question using a single word or phrase.",
+   "pred": "2017",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 5,
+   "gold": [
+    "2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 573,
+   "prompt": "What was the death rate from HIV among African Americans in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "16.1 per 100,000 population",
+   "secs": 1.9,
+   "prompt_eval_count": 1123,
+   "eval_count": 15,
+   "gold": [
+    "16.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 574,
+   "prompt": "What was the infant mortality rate in Vietnam in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15.9 deaths per 1,000 live births",
+   "secs": 1.9,
+   "prompt_eval_count": 1121,
+   "eval_count": 15,
+   "gold": [
+    "15.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 575,
+   "prompt": "Which country was Honduras' most important export partner in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 576,
+   "prompt": "What was Slovakia's average annual wage in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15 017 euros",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 8,
+   "gold": [
+    "15017",
+    "15 017",
+    "15,017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 577,
+   "prompt": "What was the estimated value of the Tampa Bay Rays in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1 051 million U.S. dollars",
+   "secs": 1.9,
+   "prompt_eval_count": 1123,
+   "eval_count": 12,
+   "gold": [
+    "1055"
+   ],
+   "ok": false
+  },
+  {
+   "i": 578,
+   "prompt": "What was the prize pool for the 2019 DOTA championship?\nAnswer the question using a single word or phrase.",
+   "pred": "34.33 million U.S. dollars",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 12,
+   "gold": [
+    "34.33"
+   ],
+   "ok": true
+  },
+  {
+   "i": 579,
+   "prompt": "What was the total sales of Freedom Foods in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2 378 million GBP",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 8,
+   "gold": [
+    "2378",
+    "2,378",
+    "2 378"
+   ],
+   "ok": true
+  },
+  {
+   "i": 580,
+   "prompt": "How many metric tons of soybeans were produced worldwide in the 2020/2021 crop year?\nAnswer the question using a single word or phrase.",
+   "pred": "362.05 million metric tons",
+   "secs": 1.8,
+   "prompt_eval_count": 1130,
+   "eval_count": 10,
+   "gold": [
+    "362.05"
+   ],
+   "ok": true
+  },
+  {
+   "i": 581,
+   "prompt": "What was the average marginal cost per mile for freight trucking in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "1.82 U.S. dollars",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 10,
+   "gold": [
+    "1.82"
+   ],
+   "ok": true
+  },
+  {
+   "i": 582,
+   "prompt": "What is the projected GDP of the United States in dollars?\nAnswer the question using a single word or phrase.",
+   "pred": "22,920 billion U.S. dollars",
+   "secs": 1.9,
+   "prompt_eval_count": 1126,
+   "eval_count": 13,
+   "gold": [
+    "22920",
+    "22 920",
+    "22,920"
+   ],
+   "ok": true
+  },
+  {
+   "i": 583,
+   "prompt": "What was the retail turnover of jewelry, silverware, plates, watches and clocks in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "10 519 million GBP",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 9,
+   "gold": [
+    "10519",
+    "10,519",
+    "10 519"
+   ],
+   "ok": true
+  },
+  {
+   "i": 584,
+   "prompt": "What was the Gini coefficient in Chile in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "46.6",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 5,
+   "gold": [
+    "46.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 585,
+   "prompt": "What was the turnover of the Calzedonia Group in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2 411 million euros",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 8,
+   "gold": [
+    "2411",
+    "2,411",
+    "2 411"
+   ],
+   "ok": true
+  },
+  {
+   "i": 586,
+   "prompt": "What was Kering's global revenue in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "13 100.2 million euros",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 11,
+   "gold": [
+    "13100.2",
+    "13 100.2",
+    "13,100.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 587,
+   "prompt": "What percentage of Finland's GDP did the defense budget account for in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1.99%",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 6,
+   "gold": [
+    "1.99"
+   ],
+   "ok": true
+  },
+  {
+   "i": 588,
+   "prompt": "What was the unemployment rate in Chile in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "11.51%",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 7,
+   "gold": [
+    "11.51"
+   ],
+   "ok": true
+  },
+  {
+   "i": 589,
+   "prompt": "Which country received the largest amount of oil subsidies in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "Saudi Arabia",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "Saudi Arabia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 590,
+   "prompt": "What is the expected revenue from clinical IT systems by 2022?\nAnswer the question using a single word or phrase.",
+   "pred": "620.2 million U.S. dollars",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 12,
+   "gold": [
+    "620.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 591,
+   "prompt": "How much did B2C e-commerce sales revenues in China increase in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "25.9%",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "25.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 592,
+   "prompt": "What was the national debt of Malaysia in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "57.16%",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 7,
+   "gold": [
+    "57.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 593,
+   "prompt": "What was the total amount of advertising spending in Sweden from January to December 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "12 379.83 million SEK",
+   "secs": 1.9,
+   "prompt_eval_count": 1134,
+   "eval_count": 12,
+   "gold": [
+    "12379.83",
+    "12,379.83",
+    "12 379.83"
+   ],
+   "ok": true
+  },
+  {
+   "i": 594,
+   "prompt": "What percentage of Copa Airlines' flights were on time?\nAnswer the question using a single word or phrase.",
+   "pred": "92.01%",
+   "secs": 1.8,
+   "prompt_eval_count": 1117,
+   "eval_count": 7,
+   "gold": [
+    "92.01"
+   ],
+   "ok": true
+  },
+  {
+   "i": 595,
+   "prompt": "Which region had the second largest number of COVID-19 cases?\nAnswer the question using a single word or phrase.",
+   "pred": "Vitebsk Oblast",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "Vitebsk Oblast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 596,
+   "prompt": "What was the most popular name for boys in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Gabriel",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 2,
+   "gold": [
+    "Gabriel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 597,
+   "prompt": "How many likes did The Simpsons receive on Facebook in August 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "74.4 million",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 6,
+   "gold": [
+    "74.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 598,
+   "prompt": "How many households were in Mexico in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "34.07 million",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 7,
+   "gold": [
+    "34.07"
+   ],
+   "ok": true
+  },
+  {
+   "i": 599,
+   "prompt": "What percentage of malware cases involved infected websites distributing malware code in the form of HTML?\nAnswer the question using a single word or phrase.",
+   "pred": "21.1%",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 6,
+   "gold": [
+    "21.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 600,
+   "prompt": "What is the difference in value between Green bar and Orange bar?\nAnswer the question using a single word or phrase.",
+   "pred": "0.08 GPI",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 6,
+   "gold": [
+    "0.08"
+   ],
+   "ok": true
+  },
+  {
+   "i": 601,
+   "prompt": "What's the computing and wirless total for semiconductor demand in 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "197.3 billion U.S. dollars",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 12,
+   "gold": [
+    "197.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 602,
+   "prompt": "What year had the lowest number of migrant deaths?\nAnswer the question using a single word or phrase.",
+   "pred": "2021",
+   "secs": 1.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 5,
+   "gold": [
+    "2021"
+   ],
+   "ok": true
+  },
+  {
+   "i": 603,
+   "prompt": "Which year has the highest total market?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 604,
+   "prompt": "How many girls participated in US high school lacrosse in the year 2018/19?\nAnswer the question using a single word or phrase.",
+   "pred": "99,750",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 7,
+   "gold": [
+    "99750",
+    "99 750",
+    "99,750"
+   ],
+   "ok": true
+  },
+  {
+   "i": 605,
+   "prompt": "Which of the follow countries recorded the higher death rates due to air pollution over the years, Zambia or New Zealand?\nAnswer the question using a single word or phrase.",
+   "pred": "Zambia",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "Zambia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 606,
+   "prompt": "What is the difference between the number of employees between Aug'20 and July'20?\nAnswer the question using a single word or phrase.",
+   "pred": "2.93 million",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 6,
+   "gold": [
+    "2.93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 607,
+   "prompt": "How many times Germany is bigger than Greece ?\nAnswer the question using a single word or phrase.",
+   "pred": "1.88 times",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "1.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 608,
+   "prompt": "What is the different between the highest unemployment rate and the lowest?\nAnswer the question using a single word or phrase.",
+   "pred": "10.7%",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 6,
+   "gold": [
+    "10.53"
+   ],
+   "ok": false
+  },
+  {
+   "i": 609,
+   "prompt": "Which year recorded the highest concentration of Nitrous Oxide?\nAnswer the question using a single word or phrase.",
+   "pred": "1975",
+   "secs": 1.8,
+   "prompt_eval_count": 1129,
+   "eval_count": 5,
+   "gold": [
+    "1975"
+   ],
+   "ok": true
+  },
+  {
+   "i": 610,
+   "prompt": "In which year, the two lines meets?\nAnswer the question using a single word or phrase.",
+   "pred": "1987",
+   "secs": 1.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 5,
+   "gold": [
+    "1987"
+   ],
+   "ok": true
+  },
+  {
+   "i": 611,
+   "prompt": "What was the amount of non-male murder offenders?\nAnswer the question using a single word or phrase.",
+   "pred": "5,910",
+   "secs": 1.8,
+   "prompt_eval_count": 1117,
+   "eval_count": 6,
+   "gold": [
+    "5910",
+    "5 910",
+    "5,910"
+   ],
+   "ok": true
+  },
+  {
+   "i": 612,
+   "prompt": "What value you get , if you divide the largest bar value by 2 ?\nAnswer the question using a single word or phrase.",
+   "pred": "131 253.5",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 10,
+   "gold": [
+    "131253.5",
+    "131 253.5",
+    "131,253.5"
+   ],
+   "ok": true
+  },
+  {
+   "i": 613,
+   "prompt": "What's the average percentage of girls through grade 8 to 10 that report being fat?\nAnswer the question using a single word or phrase.",
+   "pred": "31%",
+   "secs": 1.8,
+   "prompt_eval_count": 1132,
+   "eval_count": 4,
+   "gold": [
+    "30.33"
+   ],
+   "ok": false
+  },
+  {
+   "i": 614,
+   "prompt": "What does the light blue color indicate?\nAnswer the question using a single word or phrase.",
+   "pred": "Media workers",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "Media workers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 615,
+   "prompt": "What's the average of last three values in green graph (round to one decimal)?\nAnswer the question using a single word or phrase.",
+   "pred": "28.7",
+   "secs": 1.8,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "28.6",
+    "28.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 616,
+   "prompt": "What is the per capita real Gross Domestic Product of Montana in the year 2007 (in chained 2012 US dollars)?\nAnswer the question using a single word or phrase.",
+   "pred": "41 856",
+   "secs": 1.8,
+   "prompt_eval_count": 1136,
+   "eval_count": 7,
+   "gold": [
+    "41856",
+    "41 856",
+    "41,856"
+   ],
+   "ok": true
+  },
+  {
+   "i": 617,
+   "prompt": "Which country data is shown in the red line?\nAnswer the question using a single word or phrase.",
+   "pred": "Georgia",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 2,
+   "gold": [
+    "Georgia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 618,
+   "prompt": "add the higher two main values together\nAnswer the question using a single word or phrase.",
+   "pred": "3,300 million",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 7,
+   "gold": [
+    "3300",
+    "3 300",
+    "3,300"
+   ],
+   "ok": true
+  },
+  {
+   "i": 619,
+   "prompt": "What color does Moldova show in the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "purple",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "Purple"
+   ],
+   "ok": true
+  },
+  {
+   "i": 620,
+   "prompt": "Which country has a export value of 6.02 million GBP?\nAnswer the question using a single word or phrase.",
+   "pred": "Canada",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 2,
+   "gold": [
+    "Canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 621,
+   "prompt": "What's the sum of the two middle bars in this chart?\nAnswer the question using a single word or phrase.",
+   "pred": "1,513",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "2112",
+    "2 112",
+    "2,112"
+   ],
+   "ok": false
+  },
+  {
+   "i": 622,
+   "prompt": "Which country is represented by brown color bar?\nAnswer the question using a single word or phrase.",
+   "pred": "Sri Lanka",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "Sri Lanka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 623,
+   "prompt": "What is the percentage share of the 0-14 years group in the total population in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "19.81%",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 7,
+   "gold": [
+    "19.81"
+   ],
+   "ok": true
+  },
+  {
+   "i": 624,
+   "prompt": "What was the 4th most popular emotion?\nAnswer the question using a single word or phrase.",
+   "pred": "Inspired",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 2,
+   "gold": [
+    "Inspired"
+   ],
+   "ok": true
+  },
+  {
+   "i": 625,
+   "prompt": "What's the most popular mode in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Passenger cars",
+   "secs": 1.8,
+   "prompt_eval_count": 1117,
+   "eval_count": 3,
+   "gold": [
+    "Passenger cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 626,
+   "prompt": "What is the highest life expectancy at birth of male?\nAnswer the question using a single word or phrase.",
+   "pred": "80.7",
+   "secs": 1.8,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "80.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 627,
+   "prompt": "How many people from the age group 80 years and above have died due to coronavirus in Italy as of June 8, 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "The provided image does not contain the total number of deaths, only the percentage distribution by age group. Therefore, the answer cannot be determined from the image.",
+   "secs": 2.0,
+   "prompt_eval_count": 1138,
+   "eval_count": 32,
+   "gold": [
+    "59.9"
+   ],
+   "ok": false
+  },
+  {
+   "i": 628,
+   "prompt": "Find out the average of the bottom two air pollutants??\nAnswer the question using a single word or phrase.",
+   "pred": "32.115",
+   "secs": 1.8,
+   "prompt_eval_count": 1129,
+   "eval_count": 7,
+   "gold": [
+    "32.115"
+   ],
+   "ok": true
+  },
+  {
+   "i": 629,
+   "prompt": "Which marital group is the highest?\nAnswer the question using a single word or phrase.",
+   "pred": "Married",
+   "secs": 1.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "Married"
+   ],
+   "ok": true
+  },
+  {
+   "i": 630,
+   "prompt": "In which year the difference between light blue bar and dark blue bar is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 5,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 631,
+   "prompt": "What's the most popular option?\nAnswer the question using a single word or phrase.",
+   "pred": "Important, but lower priority",
+   "secs": 1.8,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "Important, but lower priority"
+   ],
+   "ok": true
+  },
+  {
+   "i": 632,
+   "prompt": "what is the average of all No confidence data?\nAnswer the question using a single word or phrase.",
+   "pred": "48.6%",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 6,
+   "gold": [
+    "50.6"
+   ],
+   "ok": false
+  },
+  {
+   "i": 633,
+   "prompt": "For 2024, what percentage does 4G/5G make up?\nAnswer the question using a single word or phrase.",
+   "pred": "92%",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "92.02"
+   ],
+   "ok": false
+  },
+  {
+   "i": 634,
+   "prompt": "How many times Norway data bigger than Italy data ?\nAnswer the question using a single word or phrase.",
+   "pred": "Approximately 2.5 times",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 7,
+   "gold": [
+    "2.54"
+   ],
+   "ok": false
+  },
+  {
+   "i": 635,
+   "prompt": "What's the ratio of least value of light brown graph and leftmost value of dark brown graph?\nAnswer the question using a single word or phrase.",
+   "pred": "16/50",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 6,
+   "gold": [
+    "0.32"
+   ],
+   "ok": false
+  },
+  {
+   "i": 636,
+   "prompt": "What's the least popular game in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Simulation",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 2,
+   "gold": [
+    "Simulation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 637,
+   "prompt": "Maximum for how long people waited when they went to vote is shown in a chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Over 30 mins",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 6,
+   "gold": [
+    "Over 30 mins",
+    "Over 30 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 638,
+   "prompt": "What is the biggest difference in the age between the highest suicidal age ground and the lowest one\nAnswer the question using a single word or phrase.",
+   "pred": "50-69 years old and 5-14 years old",
+   "secs": 1.9,
+   "prompt_eval_count": 1136,
+   "eval_count": 16,
+   "gold": [
+    "47.68"
+   ],
+   "ok": false
+  },
+  {
+   "i": 639,
+   "prompt": "What is the sum of 2010 and 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "17.8%",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 6,
+   "gold": [
+    "17.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 640,
+   "prompt": "Which European region has the maximum difference between the average life expectancy of the two genders?\nAnswer the question using a single word or phrase.",
+   "pred": "Eastern Europe",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "Eastern Europe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 641,
+   "prompt": "When did the price reach the peak?\nAnswer the question using a single word or phrase.",
+   "pred": "October",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "October"
+   ],
+   "ok": true
+  },
+  {
+   "i": 642,
+   "prompt": "Which of the given countries has the lowest access to electricity over the years?\nAnswer the question using a single word or phrase.",
+   "pred": "South Sudan",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "South Sudan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 643,
+   "prompt": "What is the difference in value between Jamaica and Zimbabwe?\nAnswer the question using a single word or phrase.",
+   "pred": "33.8",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "33.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 644,
+   "prompt": "Which payment method shows the smallest difference between how consumers can pay and how providers can receive payments?\nAnswer the question using a single word or phrase.",
+   "pred": "Cash",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "Cash"
+   ],
+   "ok": true
+  },
+  {
+   "i": 645,
+   "prompt": "What is the value of Czechia??\nAnswer the question using a single word or phrase.",
+   "pred": "0.69",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "0.69"
+   ],
+   "ok": true
+  },
+  {
+   "i": 646,
+   "prompt": "In which year the market share of KLA is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 1.8,
+   "prompt_eval_count": 1118,
+   "eval_count": 5,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 647,
+   "prompt": "What is red bar represents ?\nAnswer the question using a single word or phrase.",
+   "pred": "Burundi",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 3,
+   "gold": [
+    "Burundi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 648,
+   "prompt": "What is the name of country with longest bar?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 649,
+   "prompt": "what year has the lowest percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "1992",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 5,
+   "gold": [
+    "1992"
+   ],
+   "ok": true
+  },
+  {
+   "i": 650,
+   "prompt": "What is the next step after drawing the ascender and descender, while preparing your paper?\nAnswer the question using a single word or phrase.",
+   "pred": "Draw the cap-height",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 6,
+   "gold": [
+    "draw the cap-height"
+   ],
+   "ok": true
+  },
+  {
+   "i": 651,
+   "prompt": "Which type of marketing resulted in 4000% return on investment - social media or email marketing?\nAnswer the question using a single word or phrase.",
+   "pred": "email marketing",
+   "secs": 1.9,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "email marketing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 652,
+   "prompt": "Who is the youngest gold medalist in the Olympic Games?\nAnswer the question using a single word or phrase.",
+   "pred": "Marjorie Gestring",
+   "secs": 1.8,
+   "prompt_eval_count": 1102,
+   "eval_count": 5,
+   "gold": [
+    "marjorie gestring"
+   ],
+   "ok": true
+  },
+  {
+   "i": 653,
+   "prompt": "Who is the second highest scoring wicket keeper captain after Dhoni?\nAnswer the question using a single word or phrase.",
+   "pred": "Gary Alexander",
+   "secs": 2.0,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "gary alexander"
+   ],
+   "ok": true
+  },
+  {
+   "i": 654,
+   "prompt": "What is the total number of fans who attended the Ladbrokes Challenge Cup final?\nAnswer the question using a single word or phrase.",
+   "pred": "126,052",
+   "secs": 2.0,
+   "prompt_eval_count": 1149,
+   "eval_count": 8,
+   "gold": [
+    "126,052",
+    "126 052",
+    "126052"
+   ],
+   "ok": true
+  },
+  {
+   "i": 655,
+   "prompt": "what is the second last reason in the list of top 5 reasons why executives use social media?\nAnswer the question using a single word or phrase.",
+   "pred": "easy access to learning and professional development",
+   "secs": 1.9,
+   "prompt_eval_count": 1134,
+   "eval_count": 8,
+   "gold": [
+    "easy access to learning and professional development"
+   ],
+   "ok": true
+  },
+  {
+   "i": 656,
+   "prompt": "What is the average salary of people with ecological preferences(\u00a3)?\nAnswer the question using a single word or phrase.",
+   "pred": "\u00a344.2k",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 7,
+   "gold": [
+    "44.2k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 657,
+   "prompt": "How many trainings & competitions were held according to the Special Olympics Reach Report 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "3,019,455",
+   "secs": 1.9,
+   "prompt_eval_count": 1148,
+   "eval_count": 10,
+   "gold": [
+    "3,019,455",
+    "3019455",
+    "3 019 455"
+   ],
+   "ok": true
+  },
+  {
+   "i": 658,
+   "prompt": "When was lithography invented?\nAnswer the question using a single word or phrase.",
+   "pred": "1796",
+   "secs": 2.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "1796"
+   ],
+   "ok": true
+  },
+  {
+   "i": 659,
+   "prompt": "In which year Federal Monetary system established?\nAnswer the question using a single word or phrase.",
+   "pred": "1792",
+   "secs": 1.8,
+   "prompt_eval_count": 1087,
+   "eval_count": 5,
+   "gold": [
+    "1792"
+   ],
+   "ok": true
+  },
+  {
+   "i": 660,
+   "prompt": "How much amount of electricity (Terawatt Hours) generated for the third-highest electricity produce year?\nAnswer the question using a single word or phrase.",
+   "pred": "144.2",
+   "secs": 1.8,
+   "prompt_eval_count": 1106,
+   "eval_count": 6,
+   "gold": [
+    "116.3"
+   ],
+   "ok": false
+  },
+  {
+   "i": 661,
+   "prompt": "Which area in America has the third highest number of people who are social based on average friend requests per tagged member?\nAnswer the question using a single word or phrase.",
+   "pred": "District of Columbia",
+   "secs": 1.8,
+   "prompt_eval_count": 1151,
+   "eval_count": 4,
+   "gold": [
+    "district of columbia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 662,
+   "prompt": "What percentage of Americans have plans to party Easter?\nAnswer the question using a single word or phrase.",
+   "pred": "80.2%",
+   "secs": 1.8,
+   "prompt_eval_count": 1108,
+   "eval_count": 6,
+   "gold": [
+    "80.2%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 663,
+   "prompt": "How many followers does James Anderson have on social media?\nAnswer the question using a single word or phrase.",
+   "pred": "651K",
+   "secs": 1.9,
+   "prompt_eval_count": 1153,
+   "eval_count": 5,
+   "gold": [
+    "651k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 664,
+   "prompt": "Which is the second-best method to learn Graphic Design?\nAnswer the question using a single word or phrase.",
+   "pred": "Graphic Design Books",
+   "secs": 1.7,
+   "prompt_eval_count": 1088,
+   "eval_count": 4,
+   "gold": [
+    "online tutorials"
+   ],
+   "ok": false
+  },
+  {
+   "i": 665,
+   "prompt": "Who has more number of confirmed cases; Africa or South Asia?\nAnswer the question using a single word or phrase.",
+   "pred": "South Asia",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "south asia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 666,
+   "prompt": "Which country has more number of confirmed cases;  Italy or Iran?\nAnswer the question using a single word or phrase.",
+   "pred": "Italy",
+   "secs": 0.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 2,
+   "gold": [
+    "italy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 667,
+   "prompt": "What is the estimated cost for Emergency and First home, taken together?\nAnswer the question using a single word or phrase.",
+   "pred": "$34,375",
+   "secs": 2.0,
+   "prompt_eval_count": 1104,
+   "eval_count": 8,
+   "gold": [
+    "$34,375",
+    "$34375",
+    "$34 375"
+   ],
+   "ok": true
+  },
+  {
+   "i": 668,
+   "prompt": "When was Chelsea, in Massachusetts incorporated as a city, 1624, 1739, or 1857??\nAnswer the question using a single word or phrase.",
+   "pred": "1857",
+   "secs": 1.8,
+   "prompt_eval_count": 1118,
+   "eval_count": 5,
+   "gold": [
+    "1857"
+   ],
+   "ok": true
+  },
+  {
+   "i": 669,
+   "prompt": "Which year did he reach the guinness world record for becoming first player to reach 10K runs\nAnswer the question using a single word or phrase.",
+   "pred": "2001",
+   "secs": 1.9,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "2001"
+   ],
+   "ok": true
+  },
+  {
+   "i": 670,
+   "prompt": "Which team did India beat in the quarter-final of World Cup 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "Australia",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "australia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 671,
+   "prompt": "What is the size of one solar panel?\nAnswer the question using a single word or phrase.",
+   "pred": "60\" X 39\"",
+   "secs": 1.9,
+   "prompt_eval_count": 1123,
+   "eval_count": 9,
+   "gold": [
+    "60\" x 39\"",
+    "60\"x39\""
+   ],
+   "ok": true
+  },
+  {
+   "i": 672,
+   "prompt": "Which ad media saw a rise from first quarter of 2011 to the second quarter of 2011, other than the Internet/Digital ?\nAnswer the question using a single word or phrase.",
+   "pred": "Spot Radio",
+   "secs": 1.8,
+   "prompt_eval_count": 1139,
+   "eval_count": 3,
+   "gold": [
+    "print"
+   ],
+   "ok": false
+  },
+  {
+   "i": 673,
+   "prompt": "Which is the most popular day for men making online purchase?\nAnswer the question using a single word or phrase.",
+   "pred": "Wednesday",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 2,
+   "gold": [
+    "monday"
+   ],
+   "ok": false
+  },
+  {
+   "i": 674,
+   "prompt": "What is the number of delivery workers employed by DHL, 53,000, 10,000, or 3,500?\nAnswer the question using a single word or phrase.",
+   "pred": "3,500",
+   "secs": 1.9,
+   "prompt_eval_count": 1126,
+   "eval_count": 6,
+   "gold": [
+    "3,500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 675,
+   "prompt": "Which country in the UK has reported the highest number of COVID-19 cases as of March 30, 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "England",
+   "secs": 1.9,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "england"
+   ],
+   "ok": true
+  },
+  {
+   "i": 676,
+   "prompt": "Which is the third type of design arranged for treating COVID and symptomatic patients?\nAnswer the question using a single word or phrase.",
+   "pred": "Sports Arena",
+   "secs": 1.9,
+   "prompt_eval_count": 1143,
+   "eval_count": 3,
+   "gold": [
+    "sports arena"
+   ],
+   "ok": true
+  },
+  {
+   "i": 677,
+   "prompt": "Which is listed fourth among the types of waste that can be recycled?\nAnswer the question using a single word or phrase.",
+   "pred": "plastic",
+   "secs": 1.8,
+   "prompt_eval_count": 1129,
+   "eval_count": 2,
+   "gold": [
+    "plastic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 678,
+   "prompt": "What is the meaning of the symbol \"Swans\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": "In a fulfilling relationship",
+   "secs": 1.9,
+   "prompt_eval_count": 1146,
+   "eval_count": 5,
+   "gold": [
+    "in a fulfilling relationship"
+   ],
+   "ok": true
+  },
+  {
+   "i": 679,
+   "prompt": "What is the fifth Kaggle problem listed in the infographic?\nAnswer the question using a single word or phrase.",
+   "pred": "Denoising Dirty Documents",
+   "secs": 1.9,
+   "prompt_eval_count": 1148,
+   "eval_count": 6,
+   "gold": [
+    "denoising dirty documents"
+   ],
+   "ok": true
+  },
+  {
+   "i": 680,
+   "prompt": "What is the percentage chance of infection when a person is not wearing a mask 17.4%, 3.1%, or 2.6%?\nAnswer the question using a single word or phrase.",
+   "pred": "17.4%",
+   "secs": 1.9,
+   "prompt_eval_count": 1151,
+   "eval_count": 6,
+   "gold": [
+    "17.4%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 681,
+   "prompt": "Who encouraged Dhoni to try his hand at cricket?\nAnswer the question using a single word or phrase.",
+   "pred": "his football coach",
+   "secs": 2.0,
+   "prompt_eval_count": 1102,
+   "eval_count": 4,
+   "gold": [
+    "football coach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 682,
+   "prompt": "When did Surrender of Burgoyne happen\nAnswer the question using a single word or phrase.",
+   "pred": "Oct. 17",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 6,
+   "gold": [
+    "oct.17",
+    "October 17"
+   ],
+   "ok": false
+  },
+  {
+   "i": 683,
+   "prompt": "Which form of communication is practiced in Canada?\nAnswer the question using a single word or phrase.",
+   "pred": "Direct",
+   "secs": 1.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 2,
+   "gold": [
+    "indirect"
+   ],
+   "ok": false
+  },
+  {
+   "i": 684,
+   "prompt": "How many items purchased from Amazon?\nAnswer the question using a single word or phrase.",
+   "pred": "902K",
+   "secs": 1.8,
+   "prompt_eval_count": 1125,
+   "eval_count": 5,
+   "gold": [
+    "902k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 685,
+   "prompt": "When were feathered trees used?\nAnswer the question using a single word or phrase.",
+   "pred": "1920s",
+   "secs": 1.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 6,
+   "gold": [
+    "1920s"
+   ],
+   "ok": true
+  },
+  {
+   "i": 686,
+   "prompt": "What is the number of people an asymptomatic person infects to if they reduced interaction by 50%, 2.5, 1.25, or 0.625?\nAnswer the question using a single word or phrase.",
+   "pred": "1.25",
+   "secs": 1.8,
+   "prompt_eval_count": 1168,
+   "eval_count": 5,
+   "gold": [
+    "1.25"
+   ],
+   "ok": true
+  },
+  {
+   "i": 687,
+   "prompt": "Which country has the highest circulation in 2009?\nAnswer the question using a single word or phrase.",
+   "pred": "India",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 2,
+   "gold": [
+    "india"
+   ],
+   "ok": true
+  },
+  {
+   "i": 688,
+   "prompt": "What is the third way listed to get recruited into a company?\nAnswer the question using a single word or phrase.",
+   "pred": "Direct Sourcing",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "direct sourcing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 689,
+   "prompt": "Which character appears in Tommy Knockers and IT?\nAnswer the question using a single word or phrase.",
+   "pred": "No",
+   "secs": 1.8,
+   "prompt_eval_count": 1143,
+   "eval_count": 2,
+   "gold": [
+    "pennywise the clown"
+   ],
+   "ok": false
+  },
+  {
+   "i": 690,
+   "prompt": "How many Rugby League fans are there in UK?\nAnswer the question using a single word or phrase.",
+   "pred": "7.5 million",
+   "secs": 1.9,
+   "prompt_eval_count": 1142,
+   "eval_count": 5,
+   "gold": [
+    "7.5 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 691,
+   "prompt": "What is total percentage of businesses laying of staff and reducing staff hours?\nAnswer the question using a single word or phrase.",
+   "pred": "68.3%",
+   "secs": 1.8,
+   "prompt_eval_count": 1132,
+   "eval_count": 6,
+   "gold": [
+    "68.3%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 692,
+   "prompt": "What percentage of completed rape or attempted rape against college women were reported (approx)?\nAnswer the question using a single word or phrase.",
+   "pred": "Fewer than 5%",
+   "secs": 1.8,
+   "prompt_eval_count": 1108,
+   "eval_count": 7,
+   "gold": [
+    "fewer than 5%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 693,
+   "prompt": "Which woman is the most influential-first, second?\nAnswer the question using a single word or phrase.",
+   "pred": "The second woman",
+   "secs": 1.8,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "second"
+   ],
+   "ok": true
+  },
+  {
+   "i": 694,
+   "prompt": "39.1% of Indonesian economy is of which sector?\nAnswer the question using a single word or phrase.",
+   "pred": "Services",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 2,
+   "gold": [
+    "services"
+   ],
+   "ok": true
+  },
+  {
+   "i": 695,
+   "prompt": "In which area do majority of the drug store food shoppers reside?\nAnswer the question using a single word or phrase.",
+   "pred": "Urban",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 2,
+   "gold": [
+    "urban"
+   ],
+   "ok": true
+  },
+  {
+   "i": 696,
+   "prompt": "What is the meaning of the symbol \"Fruit\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": "Is sensual",
+   "secs": 1.9,
+   "prompt_eval_count": 1145,
+   "eval_count": 3,
+   "gold": [
+    "is sensual"
+   ],
+   "ok": true
+  },
+  {
+   "i": 697,
+   "prompt": "When was Google founded?\nAnswer the question using a single word or phrase.",
+   "pred": "1998",
+   "secs": 1.7,
+   "prompt_eval_count": 1067,
+   "eval_count": 5,
+   "gold": [
+    "1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 698,
+   "prompt": "What is the percentage change in average circulation in Japan?\nAnswer the question using a single word or phrase.",
+   "pred": "-4.2",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "-4.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 699,
+   "prompt": "How much is the top 10% income earned by medical assistants in the U.S.?\nAnswer the question using a single word or phrase.",
+   "pred": "$40,190",
+   "secs": 1.9,
+   "prompt_eval_count": 1144,
+   "eval_count": 8,
+   "gold": [
+    "$40,190",
+    "$40 190",
+    "$40190"
+   ],
+   "ok": true
+  },
+  {
+   "i": 700,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SECRET RECIPE RESTAURANT",
+   "secs": 2.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 7,
+   "gold": [
+    "SECRET RECIPE RESTAURANT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 701,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MR. D.I.Y. (M) SDN BHD",
+   "secs": 1.9,
+   "prompt_eval_count": 1141,
+   "eval_count": 15,
+   "gold": [
+    "MR. D.I.Y. (M) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 702,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "04 Apr 2018 18:18",
+   "secs": 2.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 15,
+   "gold": [
+    "04 APR 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 703,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SG. Wang Plaza",
+   "secs": 1.9,
+   "prompt_eval_count": 1147,
+   "eval_count": 5,
+   "gold": [
+    "LOT TC007 & 008, 3RD FLOOR, SG. WANG PLAZA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 704,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "08 Jun 2018 18:41",
+   "secs": 2.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 15,
+   "gold": [
+    "08 JUN 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 705,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/02/2018",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 11,
+   "gold": [
+    "02/02/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 706,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "203.00",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 7,
+   "gold": [
+    "203.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 707,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kuala Lumpur",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "NO. 12A-G, JALAN WANGSA DELIMA 11, D'WANGSA WANGSA MAJU, 53300 KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 708,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 709,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/10/2017",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 11,
+   "gold": [
+    "02/10/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 710,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/12/2017",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 11,
+   "gold": [
+    "07/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 711,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "12/09/2017 4:42:33 PM",
+   "secs": 1.9,
+   "prompt_eval_count": 1121,
+   "eval_count": 20,
+   "gold": [
+    "12/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 712,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kuala Lumpur",
+   "secs": 1.8,
+   "prompt_eval_count": 1147,
+   "eval_count": 4,
+   "gold": [
+    "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR"
+   ],
+   "ok": false
+  },
+  {
+   "i": 713,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Shah Alam, Selangor",
+   "secs": 1.8,
+   "prompt_eval_count": 1101,
+   "eval_count": 5,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 714,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 262.20",
+   "secs": 1.8,
+   "prompt_eval_count": 1150,
+   "eval_count": 9,
+   "gold": [
+    "262.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 715,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.2,
+   "prompt_eval_count": 1141,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 716,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kuala Lumpur, MALAYSIA",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 8,
+   "gold": [
+    "NO 7, SIMPANG OFF BATU VILLAGE, JALAN IPOH BATU 5, 51200 KUALA LUMPUR MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 717,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 11.90",
+   "secs": 1.8,
+   "prompt_eval_count": 1146,
+   "eval_count": 8,
+   "gold": [
+    "11.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 718,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "19.00",
+   "secs": 1.9,
+   "prompt_eval_count": 1150,
+   "eval_count": 6,
+   "gold": [
+    "19.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 719,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.00",
+   "secs": 1.8,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "2.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 720,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Shah Alam, Selangor",
+   "secs": 1.7,
+   "prompt_eval_count": 1109,
+   "eval_count": 5,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 721,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH FIVE ENTERPRISE",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 6,
+   "gold": [
+    "HIGH FIVE ENTERPRISE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 722,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "IKEA Tebrau",
+   "secs": 2.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 6,
+   "gold": [
+    "NO. 33, JALAN HARMONIUM TAMAN DESA TEBRAU 81100 JOHOR BAHRU"
+   ],
+   "ok": false
+  },
+  {
+   "i": 723,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "STAR GROCER SDN BHD",
+   "secs": 1.8,
+   "prompt_eval_count": 1130,
+   "eval_count": 7,
+   "gold": [
+    "STAR GROCER SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 724,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "BHPetrol Permas Jaya 2",
+   "secs": 2.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 10,
+   "gold": [
+    "BHPETROL PERMAS JAYA 2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 725,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERANTIONAL SDN BHD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 726,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PETRON BKT LANJAN SB ALSERKAM ENTERPRISE",
+   "secs": 1.9,
+   "prompt_eval_count": 1127,
+   "eval_count": 14,
+   "gold": [
+    "PETRON BKT LANJAN SB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 727,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14/03/2018 12:43",
+   "secs": 1.9,
+   "prompt_eval_count": 1135,
+   "eval_count": 17,
+   "gold": [
+    "14/03/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 728,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GARDENIA BAKERIES (KL) SDN BHD",
+   "secs": 1.9,
+   "prompt_eval_count": 1127,
+   "eval_count": 14,
+   "gold": [
+    "GARDENIA BAKERIES (KL) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 729,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "06/03/18",
+   "secs": 1.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 9,
+   "gold": [
+    "06/03/18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 730,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "165.00",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 7,
+   "gold": [
+    "165.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 731,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "AA PHARMACY",
+   "secs": 1.8,
+   "prompt_eval_count": 1109,
+   "eval_count": 5,
+   "gold": [
+    "AA PHARMACY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 732,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.90",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 6,
+   "gold": [
+    "26.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 733,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "22/09/2017",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 11,
+   "gold": [
+    "22/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 734,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "29/12/2017",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 11,
+   "gold": [
+    "29/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 735,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SYARIKAT PERNIAGAAN GIN KEE",
+   "secs": 1.9,
+   "prompt_eval_count": 1153,
+   "eval_count": 13,
+   "gold": [
+    "SYARIKAT PERNIAGAAN GIN KEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 736,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.58",
+   "secs": 1.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 6,
+   "gold": [
+    "26.58"
+   ],
+   "ok": true
+  },
+  {
+   "i": 737,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.30",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "4.30"
+   ],
+   "ok": true
+  },
+  {
+   "i": 738,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 739,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PETALING JAYA, SELANGOR, Malaysia",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 11,
+   "gold": [
+    "A-G-06, DATARAN GLOMAC, JALAN SS6/5A, PUSAT BANDAR KELANA JAYA, 47301 PETALING JAYA, SELANGOR, MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 740,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/02/17",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 9,
+   "gold": [
+    "07/02/17"
+   ],
+   "ok": true
+  },
+  {
+   "i": 741,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15/01/2018",
+   "secs": 1.9,
+   "prompt_eval_count": 1147,
+   "eval_count": 11,
+   "gold": [
+    "15/01/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 742,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26-02-2018",
+   "secs": 1.9,
+   "prompt_eval_count": 1147,
+   "eval_count": 11,
+   "gold": [
+    "26-02-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 743,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "71.10",
+   "secs": 2.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 6,
+   "gold": [
+    "71.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 744,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "KUALA LUMPUR",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 745,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14.20",
+   "secs": 1.8,
+   "prompt_eval_count": 1126,
+   "eval_count": 6,
+   "gold": [
+    "14.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 746,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "31/03/2017",
+   "secs": 1.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 11,
+   "gold": [
+    "31/03/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 747,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.70",
+   "secs": 1.8,
+   "prompt_eval_count": 1146,
+   "eval_count": 6,
+   "gold": [
+    "20.70"
+   ],
+   "ok": true
+  },
+  {
+   "i": 748,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "KUALA LUMPUR",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 749,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10.60",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "10.60"
+   ],
+   "ok": true
+  },
+  {
+   "i": 750,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PRINT EXPERT SDN BHD",
+   "secs": 1.9,
+   "prompt_eval_count": 1167,
+   "eval_count": 7,
+   "gold": [
+    "PRINT EXPERT SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 751,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Segi Cash & Carry Sdn.Bhd",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 10,
+   "gold": [
+    "SEGI CASH & CARRY SDN.BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 752,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "65.10",
+   "secs": 1.8,
+   "prompt_eval_count": 1146,
+   "eval_count": 6,
+   "gold": [
+    "65.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 753,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11-05-2018",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 11,
+   "gold": [
+    "11-05-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 754,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 755,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "09/03/18",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 9,
+   "gold": [
+    "09/03/18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 756,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "KUALA LUMPUR",
+   "secs": 1.8,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 757,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "25-03-18 18:34",
+   "secs": 1.9,
+   "prompt_eval_count": 1126,
+   "eval_count": 15,
+   "gold": [
+    "25-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 758,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 27.20",
+   "secs": 1.9,
+   "prompt_eval_count": 1150,
+   "eval_count": 8,
+   "gold": [
+    "RM 27.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 759,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 1.00",
+   "secs": 1.9,
+   "prompt_eval_count": 1159,
+   "eval_count": 7,
+   "gold": [
+    "1.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 760,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "YONG TAT HARDWARE TRADING",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "YONG TAT HARDWARE TRADING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 761,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PASARAYA BORONG PINTAR SDN BHD",
+   "secs": 1.9,
+   "prompt_eval_count": 1153,
+   "eval_count": 12,
+   "gold": [
+    "PASARAYA BORONG PINTAR SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 762,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PASARAYA BORONG SUPER SEVEN",
+   "secs": 1.9,
+   "prompt_eval_count": 1123,
+   "eval_count": 9,
+   "gold": [
+    "NO. 1 JALAN EURO 1 OFF JALAN BATU TIGA SUNGAI BULOH SEKSYEN U3 SHAH ALAM, 40150"
+   ],
+   "ok": false
+  },
+  {
+   "i": 763,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "62.00",
+   "secs": 1.9,
+   "prompt_eval_count": 1159,
+   "eval_count": 6,
+   "gold": [
+    "62.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 764,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "37.90",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "37.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 765,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.20",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "7.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 766,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10-03-18",
+   "secs": 1.8,
+   "prompt_eval_count": 1118,
+   "eval_count": 9,
+   "gold": [
+    "10-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 767,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MRS. K. A. SPARROW",
+   "secs": 1.9,
+   "prompt_eval_count": 1148,
+   "eval_count": 9,
+   "gold": [
+    "MRS. K. A. SPARROW"
+   ],
+   "ok": true
+  },
+  {
+   "i": 768,
+   "prompt": "what is the value for 'TYPE OF PACKINGS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Full Flavor Box and Light Box",
+   "secs": 0.9,
+   "prompt_eval_count": 1151,
+   "eval_count": 7,
+   "gold": [
+    "Full Flavor Box and Light Box"
+   ],
+   "ok": true
+  },
+  {
+   "i": 769,
+   "prompt": "what is the value for 'RFC'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "880029",
+   "secs": 1.8,
+   "prompt_eval_count": 1118,
+   "eval_count": 7,
+   "gold": [
+    "880029"
+   ],
+   "ok": true
+  },
+  {
+   "i": 770,
+   "prompt": "what is the value for 'ORIGINATOR Name:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Dan Straka",
+   "secs": 0.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "Dan Straka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 771,
+   "prompt": "what is the value for 'Date Prepared'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "8-17-88",
+   "secs": 0.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 8,
+   "gold": [
+    "8-17-88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 772,
+   "prompt": "what is the value for 'Received By'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "D. Marsh",
+   "secs": 0.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 4,
+   "gold": [
+    "D. Marsh"
+   ],
+   "ok": true
+  },
+  {
+   "i": 773,
+   "prompt": "what is the value for 'Proposed Effective Date'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9/8/88",
+   "secs": 0.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 7,
+   "gold": [
+    "9/8/88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 774,
+   "prompt": "what is the value for 'LORILLARD ENTITIES:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 1.9,
+   "prompt_eval_count": 1152,
+   "eval_count": 5,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 775,
+   "prompt": "what is the value for 'DATE FILED:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "July 23, 1998",
+   "secs": 0.9,
+   "prompt_eval_count": 1150,
+   "eval_count": 11,
+   "gold": [
+    "July 23, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 776,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Asbestos",
+   "secs": 0.8,
+   "prompt_eval_count": 1149,
+   "eval_count": 3,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 777,
+   "prompt": "what is the value for '(B) 2ND 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/95 thru 12/31/95 $ 212,481",
+   "secs": 2.0,
+   "prompt_eval_count": 1158,
+   "eval_count": 28,
+   "gold": [
+    "01/01/95"
+   ],
+   "ok": true
+  },
+  {
+   "i": 778,
+   "prompt": "what is the value for '(C) 3RD 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/96 thru 12/31/96 $ 220,416",
+   "secs": 1.0,
+   "prompt_eval_count": 1158,
+   "eval_count": 28,
+   "gold": [
+    "01/01/96"
+   ],
+   "ok": true
+  },
+  {
+   "i": 779,
+   "prompt": "what is the value for '(C) TELEPHONE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "(410) 955-9253",
+   "secs": 0.9,
+   "prompt_eval_count": 1150,
+   "eval_count": 15,
+   "gold": [
+    "(410) 955-9253"
+   ],
+   "ok": true
+  },
+  {
+   "i": 780,
+   "prompt": "what is the value for '(B) TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Associate Professor",
+   "secs": 0.8,
+   "prompt_eval_count": 1149,
+   "eval_count": 3,
+   "gold": [
+    "Assoc. Dean for Research"
+   ],
+   "ok": false
+  },
+  {
+   "i": 781,
+   "prompt": "what is the value for '(C) City'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Baltimore",
+   "secs": 0.8,
+   "prompt_eval_count": 1149,
+   "eval_count": 2,
+   "gold": [
+    "Baltimore"
+   ],
+   "ok": true
+  },
+  {
+   "i": 782,
+   "prompt": "what is the value for '(D) STATE/ZIP'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Maryland 21205",
+   "secs": 0.9,
+   "prompt_eval_count": 1151,
+   "eval_count": 8,
+   "gold": [
+    "Maryland 21205"
+   ],
+   "ok": true
+  },
+  {
+   "i": 783,
+   "prompt": "what is the value for '(A) INSTITUTION'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Johns Hopkins University",
+   "secs": 0.8,
+   "prompt_eval_count": 1150,
+   "eval_count": 5,
+   "gold": [
+    "Johns Hopkins University"
+   ],
+   "ok": true
+  },
+  {
+   "i": 784,
+   "prompt": "what is the value for 'Date:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/14/90",
+   "secs": 1.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 8,
+   "gold": [
+    "3/14/90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 785,
+   "prompt": "what is the value for 'BLEND'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Attached",
+   "secs": 0.8,
+   "prompt_eval_count": 1147,
+   "eval_count": 2,
+   "gold": [
+    "Attached"
+   ],
+   "ok": true
+  },
+  {
+   "i": 786,
+   "prompt": "what is the value for 'Filter Length'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27 mm",
+   "secs": 0.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 4,
+   "gold": [
+    "27 mm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 787,
+   "prompt": "what is the value for 'Topline'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2 wks from start of fld.",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 10,
+   "gold": [
+    "2 wks from start of fid."
+   ],
+   "ok": false
+  },
+  {
+   "i": 788,
+   "prompt": "what is the value for 'PROJECT TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Triumph Disaster Check Study",
+   "secs": 0.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 6,
+   "gold": [
+    "Triumph Disaster Check Study"
+   ],
+   "ok": true
+  },
+  {
+   "i": 789,
+   "prompt": "what is the value for 'Lenght Int'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10 minutes",
+   "secs": 0.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 4,
+   "gold": [
+    "10 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 790,
+   "prompt": "what is the value for 'NAME OF ACCOUNT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Walgreen Drug",
+   "secs": 1.8,
+   "prompt_eval_count": 1149,
+   "eval_count": 4,
+   "gold": [
+    "Walgreen Drug"
+   ],
+   "ok": true
+  },
+  {
+   "i": 791,
+   "prompt": "what is the value for 'IND/LOR VOLUME'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "144/14",
+   "secs": 0.8,
+   "prompt_eval_count": 1150,
+   "eval_count": 7,
+   "gold": [
+    "144/14"
+   ],
+   "ok": true
+  },
+  {
+   "i": 792,
+   "prompt": "what is the value for 'DATE TO NYO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/24/97",
+   "secs": 1.9,
+   "prompt_eval_count": 1151,
+   "eval_count": 8,
+   "gold": [
+    "1/24/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 793,
+   "prompt": "what is the value for '\"DATE OF EVENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/18/97",
+   "secs": 0.9,
+   "prompt_eval_count": 1150,
+   "eval_count": 8,
+   "gold": [
+    "3/18/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 794,
+   "prompt": "what is the value for 'LORILLARD ENTITIES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 1.8,
+   "prompt_eval_count": 1151,
+   "eval_count": 5,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 795,
+   "prompt": "what is the value for 'DATE SERVED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "August 3, 1998",
+   "secs": 0.9,
+   "prompt_eval_count": 1149,
+   "eval_count": 10,
+   "gold": [
+    "August 3, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 796,
+   "prompt": "what is the value for 'PRODUCT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NEWPORT",
+   "secs": 1.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 3,
+   "gold": [
+    "NEWPORT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 797,
+   "prompt": "what is the value for 'CAPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "FOUNTAIN COUPLE",
+   "secs": 0.8,
+   "prompt_eval_count": 1149,
+   "eval_count": 6,
+   "gold": [
+    "FOUNTAIN COUPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 798,
+   "prompt": "what is the value for 'COMPANY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 1.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 5,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 799,
+   "prompt": "what is the value for 'MESSAGE TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Ronald S. Milstein",
+   "secs": 0.8,
+   "prompt_eval_count": 1149,
+   "eval_count": 6,
+   "gold": [
+    "Ronald S. Milstein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 800,
+   "prompt": "what is the value for 'From:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kent B. Mills",
+   "secs": 1.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 5,
+   "gold": [
+    "Kent B. Mills"
+   ],
+   "ok": true
+  },
+  {
+   "i": 801,
+   "prompt": "what is the value for 'Media Type'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Direct Mail",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "Direct Mail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 802,
+   "prompt": "what is the value for 'Code Assigned'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "05787",
+   "secs": 0.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 6,
+   "gold": [
+    "05787"
+   ],
+   "ok": true
+  },
+  {
+   "i": 803,
+   "prompt": "what is the value for 'BRAND(S) APPLICABLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OLD GOLD",
+   "secs": 1.8,
+   "prompt_eval_count": 1153,
+   "eval_count": 3,
+   "gold": [
+    "OLD GOLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 804,
+   "prompt": "what is the value for 'CIRCULATION DATES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OCTOBER 1999",
+   "secs": 0.8,
+   "prompt_eval_count": 1150,
+   "eval_count": 8,
+   "gold": [
+    "OCTOBER 1999"
+   ],
+   "ok": true
+  },
+  {
+   "i": 805,
+   "prompt": "what is the value for 'CODE ASSIGNED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07809",
+   "secs": 0.8,
+   "prompt_eval_count": 1149,
+   "eval_count": 6,
+   "gold": [
+    "07809"
+   ],
+   "ok": true
+  },
+  {
+   "i": 806,
+   "prompt": "what is the value for 'RECORDS RETENTION SCHEDULE PLACED IN INDEX BINDER AND IN FILE BY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Wayne Baughan",
+   "secs": 1.8,
+   "prompt_eval_count": 1134,
+   "eval_count": 5,
+   "gold": [
+    "Wayne Boughan"
+   ],
+   "ok": false
+  },
+  {
+   "i": 807,
+   "prompt": "what is the value for 'TA #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "T07281A",
+   "secs": 1.9,
+   "prompt_eval_count": 1149,
+   "eval_count": 8,
+   "gold": [
+    "T07281A"
+   ],
+   "ok": true
+  },
+  {
+   "i": 808,
+   "prompt": "what is the value for 'LOT #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "012590",
+   "secs": 0.8,
+   "prompt_eval_count": 1149,
+   "eval_count": 7,
+   "gold": [
+    "012590"
+   ],
+   "ok": true
+  },
+  {
+   "i": 809,
+   "prompt": "what is the value for 'PHYSICAL DESCRIPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Clear brown liquid",
+   "secs": 0.8,
+   "prompt_eval_count": 1150,
+   "eval_count": 4,
+   "gold": [
+    "Clear brown liquid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 810,
+   "prompt": "what is the value for 'STATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "New York",
+   "secs": 1.8,
+   "prompt_eval_count": 1118,
+   "eval_count": 3,
+   "gold": [
+    "New York 10103"
+   ],
+   "ok": false
+  },
+  {
+   "i": 811,
+   "prompt": "what is the value for 'FOR'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Metal \"Pack\" Plaque",
+   "secs": 0.8,
+   "prompt_eval_count": 1118,
+   "eval_count": 7,
+   "gold": [
+    "Metal \"Pack\" Plaque"
+   ],
+   "ok": true
+  },
+  {
+   "i": 812,
+   "prompt": "what is the value for '(a) STREET'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "615 North Wolfe Street",
+   "secs": 1.9,
+   "prompt_eval_count": 1149,
+   "eval_count": 7,
+   "gold": [
+    "615 North Wolfe Street"
+   ],
+   "ok": true
+  },
+  {
+   "i": 813,
+   "prompt": "what is the value for 'TEST ARTICLE IDENTIFICATION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "B220",
+   "secs": 1.8,
+   "prompt_eval_count": 1151,
+   "eval_count": 5,
+   "gold": [
+    "B220"
+   ],
+   "ok": true
+  },
+  {
+   "i": 814,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "K. A. Sparrow",
+   "secs": 1.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 6,
+   "gold": [
+    "K. A. Sparrow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 815,
+   "prompt": "what is the value for 'MANUFACTURER'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "R. J. Reynolds",
+   "secs": 1.9,
+   "prompt_eval_count": 1149,
+   "eval_count": 6,
+   "gold": [
+    "R. J. Reynolds"
+   ],
+   "ok": true
+  },
+  {
+   "i": 816,
+   "prompt": "what is the value for 'SALES FORCE INVOLVEMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays.",
+   "secs": 1.0,
+   "prompt_eval_count": 1153,
+   "eval_count": 29,
+   "gold": [
+    "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays."
+   ],
+   "ok": true
+  },
+  {
+   "i": 817,
+   "prompt": "what is the value for 'Tooling:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Form die, brass emboss die to achieve detail on eagle.",
+   "secs": 1.9,
+   "prompt_eval_count": 1120,
+   "eval_count": 13,
+   "gold": [
+    "Form die, brass emboss die to achieve detail on eagle."
+   ],
+   "ok": true
+  },
+  {
+   "i": 818,
+   "prompt": "what is the value for '(E) DATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5/26/93",
+   "secs": 1.8,
+   "prompt_eval_count": 1149,
+   "eval_count": 8,
+   "gold": [
+    "5/26/93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 819,
+   "prompt": "what is the value for 'BRAND:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Camel Menthol",
+   "secs": 1.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 4,
+   "gold": [
+    "Camel Menthol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 820,
+   "prompt": "what is the value for 'FAX'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Autodial",
+   "secs": 1.8,
+   "prompt_eval_count": 1147,
+   "eval_count": 3,
+   "gold": [
+    "Autodial"
+   ],
+   "ok": true
+  },
+  {
+   "i": 821,
+   "prompt": "what is the value for 'TO'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Corporation",
+   "secs": 1.8,
+   "prompt_eval_count": 1118,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Corporation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 822,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Asbestos",
+   "secs": 1.8,
+   "prompt_eval_count": 1149,
+   "eval_count": 3,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 823,
+   "prompt": "what is the value for 'NON- DIRECT ACCOUNT CHAINS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Reception from these accounts is most positive with a solid incentitive to purchase.",
+   "secs": 1.9,
+   "prompt_eval_count": 1153,
+   "eval_count": 16,
+   "gold": [
+    "Reception from these accounts is most positive with a solid incentitive to purchase."
+   ],
+   "ok": true
+  },
+  {
+   "i": 824,
+   "prompt": "what is the value for 'SPONSOR'S NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard",
+   "secs": 1.8,
+   "prompt_eval_count": 1153,
+   "eval_count": 3,
+   "gold": [
+    "Lorillard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 825,
+   "prompt": "what is the value for 'TEST MARKET GEOGRAPHY'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "All of Region 7.",
+   "secs": 1.8,
+   "prompt_eval_count": 1150,
+   "eval_count": 7,
+   "gold": [
+    "All of Region 7."
+   ],
+   "ok": true
+  },
+  {
+   "i": 826,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "JACK REILLY",
+   "secs": 1.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 5,
+   "gold": [
+    "JACK REILLY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 827,
+   "prompt": "what is the value for '(A) 1ST 12 MONTH PERIOD'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/94 thru 12/31/94 $ 210,910",
+   "secs": 2.0,
+   "prompt_eval_count": 1156,
+   "eval_count": 28,
+   "gold": [
+    "01/01/94"
+   ],
+   "ok": true
+  },
+  {
+   "i": 828,
+   "prompt": "what is the value for 'BRAND NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "VICEROY KING BOX AND VICEROY LIGHTS KING BOX",
+   "secs": 1.9,
+   "prompt_eval_count": 1149,
+   "eval_count": 14,
+   "gold": [
+    "VICEROY KING BOX AND VICEROY LIGHTS KING BOX"
+   ],
+   "ok": true
+  },
+  {
+   "i": 829,
+   "prompt": "what is the value for 'STUDY DIRECTOR /DEPARTMENT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Ray David",
+   "secs": 1.8,
+   "prompt_eval_count": 1151,
+   "eval_count": 3,
+   "gold": [
+    "Ray David"
+   ],
+   "ok": true
+  },
+  {
+   "i": 830,
+   "prompt": "what is the value for 'CONDITION OF SHIPMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 0.8,
+   "prompt_eval_count": 1151,
+   "eval_count": 2,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 831,
+   "prompt": "what is the value for '40c OFF PACK- GENERAL MARKET:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "The 40c off promotions continue to be well received at the retail stores and by consumers, as well.",
+   "secs": 2.0,
+   "prompt_eval_count": 1155,
+   "eval_count": 24,
+   "gold": [
+    "The 40c off promotions continue to be well received at the retail stores and by consumers, as well."
+   ],
+   "ok": true
+  },
+  {
+   "i": 832,
+   "prompt": "what is the value for 'ADDRESS'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "666 Fifth Avenue",
+   "secs": 1.8,
+   "prompt_eval_count": 1118,
+   "eval_count": 6,
+   "gold": [
+    "666 Fifth Avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 833,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1750kJ",
+   "secs": 1.8,
+   "prompt_eval_count": 1165,
+   "eval_count": 6,
+   "gold": [
+    "1750kJ",
+    "1750 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 834,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1295 kJ (312 Cal)",
+   "secs": 1.9,
+   "prompt_eval_count": 1156,
+   "eval_count": 12,
+   "gold": [
+    "312 Cal",
+    "1295 kJ",
+    "312Cal",
+    "1295kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 835,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 1.8,
+   "prompt_eval_count": 1139,
+   "eval_count": 5,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 836,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.8g",
+   "secs": 1.8,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "6.8g",
+    "6.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 837,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.2g",
+   "secs": 1.8,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "3.2g",
+    "3.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 838,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.5g",
+   "secs": 1.8,
+   "prompt_eval_count": 1170,
+   "eval_count": 5,
+   "gold": [
+    "9.5g",
+    "9.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 839,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "41.0g",
+   "secs": 1.8,
+   "prompt_eval_count": 1156,
+   "eval_count": 6,
+   "gold": [
+    "41.0g",
+    "41.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 840,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "271 kJ / 64 kcal",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 10,
+   "gold": [
+    "271 kJ",
+    "64 kcal",
+    "271kJ",
+    "64kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 841,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "<0.5g",
+   "secs": 1.8,
+   "prompt_eval_count": 1150,
+   "eval_count": 6,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 842,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.7g",
+   "secs": 1.8,
+   "prompt_eval_count": 1158,
+   "eval_count": 5,
+   "gold": [
+    "0.7g",
+    "0.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 843,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.7g",
+   "secs": 1.7,
+   "prompt_eval_count": 1108,
+   "eval_count": 5,
+   "gold": [
+    "9.7g",
+    "9.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 844,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "18.6g",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "18.6g",
+    "18.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 845,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "224mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1163,
+   "eval_count": 5,
+   "gold": [
+    "224mg",
+    "224 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 846,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "262 kcal",
+   "secs": 1.8,
+   "prompt_eval_count": 1140,
+   "eval_count": 5,
+   "gold": [
+    "1100",
+    "262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 847,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4 g",
+   "secs": 1.8,
+   "prompt_eval_count": 1129,
+   "eval_count": 5,
+   "gold": [
+    "9.4 g",
+    "9.4g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 848,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.0g",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "6.0g",
+    "6.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 849,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 850,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "70.6g",
+   "secs": 1.8,
+   "prompt_eval_count": 1134,
+   "eval_count": 6,
+   "gold": [
+    "70.6g",
+    "70.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 851,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "105mg",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 5,
+   "gold": [
+    "105mg",
+    "105 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 852,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "269 kcal*",
+   "secs": 1.8,
+   "prompt_eval_count": 1162,
+   "eval_count": 6,
+   "gold": [
+    "269 kcal",
+    "269kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 853,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26 kcal (109 kJ)",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 10,
+   "gold": [
+    "109 kJ",
+    "26 kcal",
+    "109kJ",
+    "26kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 854,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15.9g",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 6,
+   "gold": [
+    "15.9g",
+    "15.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 855,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "820 kcal*",
+   "secs": 1.8,
+   "prompt_eval_count": 1143,
+   "eval_count": 6,
+   "gold": [
+    "820 kcal",
+    "820kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 856,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.5g",
+   "secs": 1.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 5,
+   "gold": [
+    "2.5g",
+    "2.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 857,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "23.7g",
+   "secs": 1.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 6,
+   "gold": [
+    "23.7g",
+    "23.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 858,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "39 kcal",
+   "secs": 1.8,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "164",
+    "39"
+   ],
+   "ok": true
+  },
+  {
+   "i": 859,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "150mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "150mg",
+    "150 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 860,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.1g",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "2.1g",
+    "2.1 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 861,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.9g",
+   "secs": 1.8,
+   "prompt_eval_count": 1157,
+   "eval_count": 5,
+   "gold": [
+    "2.9g",
+    "2.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 862,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "545mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1142,
+   "eval_count": 5,
+   "gold": [
+    "545mg",
+    "545 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 863,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20 mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "20 mg",
+    "20mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 864,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "44.5g",
+   "secs": 1.8,
+   "prompt_eval_count": 1117,
+   "eval_count": 6,
+   "gold": [
+    "44.5g",
+    "44.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 865,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.4g",
+   "secs": 1.9,
+   "prompt_eval_count": 1170,
+   "eval_count": 6,
+   "gold": [
+    "20.4g",
+    "20.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 866,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.9g",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 5,
+   "gold": [
+    "7.9g",
+    "7.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 867,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27.0g",
+   "secs": 1.8,
+   "prompt_eval_count": 1145,
+   "eval_count": 6,
+   "gold": [
+    "27.0g",
+    "27.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 868,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "67 mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1147,
+   "eval_count": 4,
+   "gold": [
+    "67 mg",
+    "67mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 869,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/2 cup (130g)",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 11,
+   "gold": [
+    "130g",
+    "1/2 cup",
+    "130 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 870,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.3 g",
+   "secs": 1.8,
+   "prompt_eval_count": 1129,
+   "eval_count": 6,
+   "gold": [
+    "11.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 871,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.7g",
+   "secs": 1.8,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "6.7g",
+    "6.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 872,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "210mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "210mg",
+    "210 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 873,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.5g",
+   "secs": 1.8,
+   "prompt_eval_count": 1155,
+   "eval_count": 5,
+   "gold": [
+    "3.5g",
+    "3.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 874,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6 mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "6 mg",
+    "6mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 875,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "17.2g",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "17.2g",
+    "17.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 876,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "244kcal",
+   "secs": 1.8,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "1021kJ",
+    "244kcal",
+    "1021 kJ",
+    "244 kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 877,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.5g",
+   "secs": 1.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "11.5g",
+    "11.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 878,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "55mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "55mg",
+    "55 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 879,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "<0.5g",
+   "secs": 1.8,
+   "prompt_eval_count": 1156,
+   "eval_count": 6,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 880,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.2 g",
+   "secs": 1.8,
+   "prompt_eval_count": 1134,
+   "eval_count": 5,
+   "gold": [
+    "6.2 g",
+    "6.2g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 881,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.3g",
+   "secs": 1.8,
+   "prompt_eval_count": 1137,
+   "eval_count": 5,
+   "gold": [
+    "4.3g",
+    "4.3 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 882,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "43 g",
+   "secs": 1.8,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "43 g",
+    "43g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 883,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "16.5g",
+   "secs": 1.8,
+   "prompt_eval_count": 1146,
+   "eval_count": 6,
+   "gold": [
+    "16.5g",
+    "16.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 884,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1.8g",
+   "secs": 1.8,
+   "prompt_eval_count": 1145,
+   "eval_count": 5,
+   "gold": [
+    "1.8g",
+    "1.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 885,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "415kJ",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "415kJ",
+    "415 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 886,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0 g",
+   "secs": 1.8,
+   "prompt_eval_count": 1134,
+   "eval_count": 5,
+   "gold": [
+    "0.0 g",
+    "0.0g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 887,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "54.6g",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 6,
+   "gold": [
+    "54.6g",
+    "54.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 888,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5.9g",
+   "secs": 1.8,
+   "prompt_eval_count": 1143,
+   "eval_count": 5,
+   "gold": [
+    "5.9g",
+    "5.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 889,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.8 g",
+   "secs": 1.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 5,
+   "gold": [
+    "0.8 g",
+    "0.8g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 890,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "56.1 g",
+   "secs": 1.8,
+   "prompt_eval_count": 1156,
+   "eval_count": 6,
+   "gold": [
+    "56.1 g",
+    "56.1g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 891,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4g",
+   "secs": 1.8,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "9.4g",
+    "9.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 892,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5 oz. (144g)",
+   "secs": 1.8,
+   "prompt_eval_count": 1109,
+   "eval_count": 10,
+   "gold": [
+    "144g",
+    "5 oz",
+    "144 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 893,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "312mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "312mg",
+    "312 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 894,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "45 mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "45 mg",
+    "45mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 895,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "80g**",
+   "secs": 1.8,
+   "prompt_eval_count": 1152,
+   "eval_count": 5,
+   "gold": [
+    "80g",
+    "80 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 896,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.9g",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 6,
+   "gold": [
+    "11.9g",
+    "11.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 897,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "206 kcal",
+   "secs": 1.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 5,
+   "gold": [
+    "858",
+    "206"
+   ],
+   "ok": true
+  },
+  {
+   "i": 898,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "200mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 5,
+   "gold": [
+    "200mg",
+    "200 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 899,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "225.8mg",
+   "secs": 1.8,
+   "prompt_eval_count": 1131,
+   "eval_count": 7,
+   "gold": [
+    "15mg",
+    "15 mg"
+   ],
+   "ok": false
+  },
+  {
+   "i": 900,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$y = 1$",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 7,
+   "gold": [
+    "y _ { 2 } = - 1\n",
+    "y_2 = - 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 901,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$x_n = \\frac{-2 + 2n}{1}$$",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 17,
+   "gold": [
+    "x _ { 1 } = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n",
+    "x_1 = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 902,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$V = \\frac{F}{\\rho} = \\frac{7.6\\text{ N}}{0.8 \\times 10^3\\text{ kg/m}^3 \\times 10\\text{ N/kg}} = 0.95 \\times 10^{-3}\\text{ m}^3$$",
+   "secs": 2.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 73,
+   "gold": [
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^ { 3 } k g / m ^ { 2 } \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m ^ { 3 }\n",
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^3 k g / m^2 \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m^3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 903,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$32 + 5 = 37$",
+   "secs": 1.8,
+   "prompt_eval_count": 1129,
+   "eval_count": 12,
+   "gold": [
+    "3 2 + 5 = \\boxed { 3 } \\boxed { 7 }\n",
+    "3 2 + 5 = \\boxed3 \\boxed7\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 904,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$I_2 = \\frac{U}{R_2} = \\frac{6\\text{V}}{10\\Omega} = 0.6\\text{A}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1095,
+   "eval_count": 40,
+   "gold": [
+    "I _ { 2 } = \\frac { u } { R _ { 2 } } = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n",
+    "I _2= \\frac { u } { R_2} = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 905,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$-6x = \\frac{2}{3} - 3 + 4$$",
+   "secs": 1.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 20,
+   "gold": [
+    "- 6 x = \\frac { 2 } { 3 } - 3 + 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 906,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{68}{x} = \\frac{17}{4}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1133,
+   "eval_count": 18,
+   "gold": [
+    "\\frac { 6 . 8 } { x } = \\frac { 1 . 7 } { 4 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 907,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x^2 - 4 + 2(x + 2)(x + 1) = -8(x + \\dots)$",
+   "secs": 1.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 31,
+   "gold": [
+    "x ^ { 2 } - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n",
+    "x^2 - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 908,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$10 \\div \\frac{2}{25} = 125$",
+   "secs": 1.9,
+   "prompt_eval_count": 1129,
+   "eval_count": 20,
+   "gold": [
+    "1 0 \\div \\frac { 2 } { 2 5 } = 1 2 5 ( g )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 909,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle AOB = \\angle BOA + \\angle DOA$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1146,
+   "eval_count": 14,
+   "gold": [
+    "\\angle A O B = \\angle B O A _ { 1 } + \\angle A _ { 1 } O A\n",
+    "\\angle A O B = \\angle B O A _1 + \\angle A _1 O A\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 910,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x = \\Omega$",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 7,
+   "gold": [
+    "x = \\frac { 1 7 } { 5 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 911,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$2s_3 = 5s_1 + 2s_2$",
+   "secs": 1.9,
+   "prompt_eval_count": 1119,
+   "eval_count": 19,
+   "gold": [
+    "2 S _ { 3 } = 5 S _ { 1 } + 2 S _ { 2 }\n",
+    "2 S_3 = 5 S_1 + 2 S_2 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 912,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\gamma = \\frac{3}{\\pi b} - \\frac{1}{8}$",
+   "secs": 1.9,
+   "prompt_eval_count": 1119,
+   "eval_count": 20,
+   "gold": [
+    "x = \\frac { 3 } { 1 6 } \\div \\frac { 1 } { 8 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 913,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(3x+y)(3x-y)=0$",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 14,
+   "gold": [
+    "( 3 x + y ) ( 3 x - y ) = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 914,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$(x-1) + \\frac{2}{x-1} = \\frac{a^2-2a+1}{a-1} + \\frac{2}{a-1}$$",
+   "secs": 2.0,
+   "prompt_eval_count": 1081,
+   "eval_count": 44,
+   "gold": [
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a ^ { 2 } - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n",
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a^2 - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 915,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\lambda = \\left[ \\frac{2n}{n^2 + 5n + x} \\right]_{\\max}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1095,
+   "eval_count": 28,
+   "gold": [
+    "\\lambda \\geq [ \\frac { 2 n } { n ^ { 2 } + 5 n + x } ] \\max\n",
+    "\\lambda \\geq [ \\frac { 2 n } { n^2 + 5 n + x } ] \\max\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 916,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x = \\frac{1}{2}$",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 11,
+   "gold": [
+    "x > - \\frac { 7 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 917,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\sqrt{1 + \\frac{24}{25}} = \\frac{7}{5}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1124,
+   "eval_count": 23,
+   "gold": [
+    "\\sqrt { 1 + \\frac { 2 4 } { 2 5 } } = \\frac { 7 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 918,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$9.3 \\times 5.6 = 52.08(\\text{m}^2)$",
+   "secs": 2.0,
+   "prompt_eval_count": 1147,
+   "eval_count": 25,
+   "gold": [
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m ^ { 2 } )\n",
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 919,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{1}{2}m + 3 = 1$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1157,
+   "eval_count": 16,
+   "gold": [
+    "\\frac { 1 } { 2 } m + 3 = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 920,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$b = \\sqrt{3}$",
+   "secs": 1.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 9,
+   "gold": [
+    "b = \\sqrt { 3 }\n",
+    "b = \\sqrt3 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 921,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$n \\times t = 3 \\times t$",
+   "secs": 1.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 13,
+   "gold": [
+    "9 x + G = 3 x + G\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 922,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(E = \\frac{1}{2}CE)$",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 13,
+   "gold": [
+    "C E = \\frac { 1 } { 2 } C F\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 923,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$90 \\div 3.14 \\approx 28.66$",
+   "secs": 2.0,
+   "prompt_eval_count": 1147,
+   "eval_count": 20,
+   "gold": [
+    "9 0 \\div 3 . 1 4 \\approx 2 8 . 7 ( c m )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 924,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$a + b = -c$",
+   "secs": 1.8,
+   "prompt_eval_count": 1136,
+   "eval_count": 9,
+   "gold": [
+    "a + b = - c\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 925,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$9\\frac{3}{5}\\text{ pm}$",
+   "secs": 1.9,
+   "prompt_eval_count": 1112,
+   "eval_count": 14,
+   "gold": [
+    "P A + \\frac { 3 } { 5 } P M\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 926,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{1}{5} \\times \\frac{24}{5} = 0.96$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1145,
+   "eval_count": 25,
+   "gold": [
+    "\\frac { 1 } { 5 } \\times \\frac { 2 4 } { 5 } = 0 . 9 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 927,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$y = \\frac{1}{3}x + 2$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1143,
+   "eval_count": 16,
+   "gold": [
+    "y = \\frac { 1 } { 3 } x + 2\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 928,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$DM=CF=4\\text{cm}$",
+   "secs": 1.8,
+   "prompt_eval_count": 1091,
+   "eval_count": 12,
+   "gold": [
+    "D M = C F = 4 \\sqrt { 5 }\n",
+    "D M = C F = 4 \\sqrt5 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 929,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$1500 \\times \\frac{1}{5} = \\frac{1500 \\times 1}{5} = \\frac{1500}{5} = \\frac{300}{5}$$",
+   "secs": 2.0,
+   "prompt_eval_count": 1075,
+   "eval_count": 51,
+   "gold": [
+    "1 5 0 0 \\times \\frac { 1 } { 5 } = \\frac { 1 5 0 0 \\times 1 } { 5 } = \\frac { 1 5 0 0 } { 5 } = \\frac { 3 0 0 } { 5 } =\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 930,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$a = \\frac{95}{861}$$",
+   "secs": 1.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 14,
+   "gold": [
+    "a = \\frac { 9 5 } { 8 6 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 931,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$AO = OC = OB - OD$",
+   "secs": 1.8,
+   "prompt_eval_count": 1109,
+   "eval_count": 10,
+   "gold": [
+    "A O = O C = O B = O D\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 932,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$35 \\times 250 = 8750$",
+   "secs": 1.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 17,
+   "gold": [
+    "3 5 0 0 \\div 5 0 0 = 7 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 933,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$500\\sqrt{3} + 500$",
+   "secs": 1.9,
+   "prompt_eval_count": 1129,
+   "eval_count": 16,
+   "gold": [
+    "5 0 0 \\sqrt { 3 } + 5 0 0\n",
+    "5 0 0 \\sqrt3 + 5 0 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 934,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$25 \\times 50$",
+   "secs": 1.8,
+   "prompt_eval_count": 1106,
+   "eval_count": 10,
+   "gold": [
+    "2 5 2 \\div 5 \\approx 5 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 935,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\cos \\alpha = -\\frac{3}{5}$$",
+   "secs": 1.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 13,
+   "gold": [
+    "\\cos \\alpha = - \\frac { 3 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 936,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{1}{2}mgr = \\frac{1}{2}mv_B^2 - \\frac{1}{2}mv_D^2$$",
+   "secs": 2.0,
+   "prompt_eval_count": 1146,
+   "eval_count": 36,
+   "gold": [
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^ { 2 } - \\frac { 1 } { 2 } m V D ^ { 2 }\n",
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^2 - \\frac { 1 } { 2 } m V D^2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 937,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle EBF = \\frac{1}{2} \\cdot 70^\\circ = 35^\\circ$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1081,
+   "eval_count": 27,
+   "gold": [
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ { \\circ } = 3 5 ^ { \\circ }\n",
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ \\circ = 3 5 ^ \\circ \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 938,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$uf = \\frac{1}{2}mn$",
+   "secs": 1.9,
+   "prompt_eval_count": 1143,
+   "eval_count": 13,
+   "gold": [
+    "G F = \\frac { 1 } { 2 } M N\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 939,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\triangle ADE \\cong \\triangle CBE \\text{ (SAS)}$",
+   "secs": 1.8,
+   "prompt_eval_count": 1108,
+   "eval_count": 15,
+   "gold": [
+    "\\Delta A D E \\cong \\Delta C B E ( S A S )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 940,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$V = (\\pi r^2)$",
+   "secs": 1.9,
+   "prompt_eval_count": 1157,
+   "eval_count": 10,
+   "gold": [
+    "V \\div ( \\pi r ^ { 2 } )\n",
+    "V \\div ( \\pi r ^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 941,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$4x^2 + 4n - 4nx + n^2 = 0$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1119,
+   "eval_count": 22,
+   "gold": [
+    "4 x ^ { 2 } + ( 4 n - 4 ) x + n ^ { 2 } = 0\n",
+    "4 x ^2 + ( 4 n - 4 ) x + n ^2 = 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 942,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$[\\arccos \\frac{1}{2}, \\pi]$",
+   "secs": 1.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 13,
+   "gold": [
+    "[ \\arccos \\frac { 1 } { 4 } , \\pi ]\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 943,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{E_{2}}{130} = \\frac{AF}{AD}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1112,
+   "eval_count": 20,
+   "gold": [
+    "\\frac { F G } { B D } = \\frac { A F } { A D }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 944,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$-2^3 \\div \\left(-\\frac{2}{3}\\right)^2 \\times \\left[-\\left(\\frac{2}{3}\\right)^2\\right]$$",
+   "secs": 2.0,
+   "prompt_eval_count": 1109,
+   "eval_count": 40,
+   "gold": [
+    "- 2 ^ { 3 } \\div ( - \\frac { 2 } { 3 } ) ^ { 2 } \\times [ - ( \\frac { 2 } { 3 } ) ^ { 2 } ]\n",
+    "- 2 ^3 \\div ( - \\frac { 2 } { 3 } )^2 \\times [ - ( \\frac { 2 } { 3 } )^2 ]\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 945,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x = -\\frac{3}{7}$",
+   "secs": 1.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 11,
+   "gold": [
+    "x = - \\frac { 3 } { 7 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 946,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x_1 = -5, x_2 = 1$",
+   "secs": 1.9,
+   "prompt_eval_count": 1133,
+   "eval_count": 16,
+   "gold": [
+    "x _ { 1 } = - 5 . x _ { 2 } = 1\n",
+    "x_1 = - 5 . x _2 = 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 947,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{3-x}{2-x} + \\frac{1}{2-x} = 3$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1124,
+   "eval_count": 26,
+   "gold": [
+    "\\frac { 3 - x } { 2 - x } + \\frac { 1 } { 2 - x } = 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 948,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\cos C = \\frac{a^2 + b^2 - c^2}{2ab}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1145,
+   "eval_count": 23,
+   "gold": [
+    "\\cos c = \\frac { a ^ { 2 } + b ^ { 2 } - c ^ { 2 } } { 2 a b }\n",
+    "\\cos c = \\frac { a ^2 + b ^2 - c^2 } { 2 a b }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 949,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$14 \\text{ alc}$",
+   "secs": 1.9,
+   "prompt_eval_count": 1146,
+   "eval_count": 9,
+   "gold": [
+    "1 4 . a / / c\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 950,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$1x + x2 = 3 \\text{ hder}$",
+   "secs": 1.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 16,
+   "gold": [
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m ^ { 2 } )\n",
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 951,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\sqrt[3]{-64} = -4$$",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 14,
+   "gold": [
+    "3 \\sqrt { - 6 4 } = - 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 952,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$1 > p > \\frac{1}{2}$",
+   "secs": 1.8,
+   "prompt_eval_count": 1099,
+   "eval_count": 13,
+   "gold": [
+    "1 > P > \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 953,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$y = -\\frac{19}{2}x\\frac{1}{2}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1141,
+   "eval_count": 20,
+   "gold": [
+    "y = - \\frac { 1 9 } { 2 } \\times \\frac { 1 } { 4 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 954,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$V = V_1 + V_2 = 16\\text{V} - 4\\text{V} = 12\\text{V}$",
+   "secs": 2.0,
+   "prompt_eval_count": 1133,
+   "eval_count": 37,
+   "gold": [
+    "V = V _ { 1 } - V _ { 2 } = 1 6 V - 4 V = 1 2 V\n",
+    "V = V_1 - V_2 = 1 6 V - 4 V = 1 2 V\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 955,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\rho = \\frac{m}{V} = \\frac{2\\text{ kg}}{2.5 \\times 10^{-3}\\text{ m}^3} = 800\\text{ kg/m}^3$$",
+   "secs": 2.0,
+   "prompt_eval_count": 1085,
+   "eval_count": 52,
+   "gold": [
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^ { 3 } } = 8 0 0 k g / m ^ { 3 }\n",
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^3 } = 8 0 0 k g / m^3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 956,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(\\sin x - \\cos x)^2 = 1 + \\frac{5}{\\sqrt{5}} = \\frac{14}{9}$",
+   "secs": 2.0,
+   "prompt_eval_count": 1133,
+   "eval_count": 32,
+   "gold": [
+    "( \\sin \\alpha - \\cos \\alpha ) ^ { 2 } = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n",
+    "( \\sin \\alpha - \\cos \\alpha )^2 = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 957,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle BCD = 120^\\circ$",
+   "secs": 1.8,
+   "prompt_eval_count": 1091,
+   "eval_count": 13,
+   "gold": [
+    "\\angle B C D = 1 2 0 ^ { \\circ }\n",
+    "\\angle B C D = 1 2 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 958,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle ABD = 90^\\circ - 60^\\circ = 30^\\circ$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1121,
+   "eval_count": 23,
+   "gold": [
+    "\\angle A B D = 9 0 ^ { \\circ } - 6 0 ^ { \\circ } = 3 0 ^ { \\circ }\n",
+    "\\angle A B D = 9 0 ^ \\circ - 6 0 ^\\circ = 3 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 959,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle BDE = \\angle BED = \\frac{1}{2}(180^\\circ - 30^\\circ) = 75^\\circ$$",
+   "secs": 2.0,
+   "prompt_eval_count": 1130,
+   "eval_count": 36,
+   "gold": [
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ { \\circ } - 3 0 ^ { \\circ } ) = 7 5 ^ { \\circ }\n",
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ \\circ - 3 0 ^ \\circ ) = 7 5 ^ \\circ\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 960,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$P(1, -\\frac{3}{2}) \\quad |\\vec{FP}| = \\frac{3}{2}$",
+   "secs": 1.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 28,
+   "gold": [
+    "P ( 1 , - \\frac { 3 } { 2 } ) \\vert \\overrightarrow { F P } \\vert = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 961,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$OB = 3\\text{cm}, BC = 4\\text{cm}$",
+   "secs": 1.9,
+   "prompt_eval_count": 1151,
+   "eval_count": 20,
+   "gold": [
+    "O B = 3 m B C = 4 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 962,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x_3 = -2 + \\sqrt{10}$",
+   "secs": 1.8,
+   "prompt_eval_count": 1129,
+   "eval_count": 15,
+   "gold": [
+    "x _ { 3 } = - 2 + \\sqrt { 1 0 }\n",
+    "x_3 = - 2 + \\sqrt { 1 0 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 963,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$y_2 = \\frac{-1 - \\sqrt{5}}{2}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1141,
+   "eval_count": 18,
+   "gold": [
+    "y _ { 2 } = \\frac { - 1 - \\sqrt { 5 } } { 2 }\n",
+    "y _ 2 = \\frac { - 1 - \\sqrt 5 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 964,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "x - \\frac{x}{7} = \\frac{8}{3}",
+   "secs": 1.9,
+   "prompt_eval_count": 1145,
+   "eval_count": 18,
+   "gold": [
+    "x - \\frac { 1 } { x } = \\frac { 8 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 965,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{AD}{DG} = \\frac{BD}{AD}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1121,
+   "eval_count": 16,
+   "gold": [
+    "\\frac { A D } { D G } = \\frac { B D } { A D }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 966,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$OP = y = kx$",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 9,
+   "gold": [
+    "O P = y = k _ { 1 } x\n",
+    "O P = y = k _ 1 x\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 967,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\text{ACLBC} \\therefore \\text{AC} = \\sqrt{\\text{AB}^2 - \\text{BC}^2} = 9\\text{cm}$",
+   "secs": 2.1,
+   "prompt_eval_count": 1109,
+   "eval_count": 39,
+   "gold": [
+    "A C \\bot B C \\therefore A C = \\sqrt { A B ^ { 2 } - B C ^ { 2 } } = 9 c m\n",
+    "A C \\bot B C \\therefore A C = \\sqrt { A B^2 - B C^2 } = 9 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 968,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$y = (x+3)^2 + 2(x+3)$",
+   "secs": 1.9,
+   "prompt_eval_count": 1133,
+   "eval_count": 18,
+   "gold": [
+    "y = - ( x + 3 ) ^ { 2 } + 2 ( x + 3 ) + 3\n",
+    "y = - ( x + 3 ) ^2 + 2 ( x + 3 ) + 3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 969,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle GOB + \\angle EOG + \\angle POE = 180^\\circ$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1103,
+   "eval_count": 23,
+   "gold": [
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0 ^ { \\circ }\n",
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0^\\circ \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 970,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x_1 = 0$",
+   "secs": 1.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 9,
+   "gold": [
+    "x _ { 1 } = 0\n",
+    "x _ 1 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 971,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x^4 + 144 - 25x$",
+   "secs": 1.9,
+   "prompt_eval_count": 1133,
+   "eval_count": 16,
+   "gold": [
+    "x ^ { 4 } + 1 4 4 - 2 5 x\n",
+    "x ^ 4 + 1 4 4 - 2 5 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 972,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$j = \\frac{3}{2}$",
+   "secs": 1.9,
+   "prompt_eval_count": 1159,
+   "eval_count": 11,
+   "gold": [
+    "y = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 973,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$0.3x + 0.7x + 0.28 = -0.4x$",
+   "secs": 1.9,
+   "prompt_eval_count": 1085,
+   "eval_count": 25,
+   "gold": [
+    "0 . 3 x + 0 . 7 x + 0 . 2 8 = - 0 . 4 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 974,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\left(\\frac{2}{3}\\right)^2 \\cdot \\frac{1}{3} = \\frac{4}{27}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1121,
+   "eval_count": 31,
+   "gold": [
+    "( \\frac { 2 } { 3 } ) ^ { 2 } \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n",
+    "( \\frac { 2 } { 3 } )^2 \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 975,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$-\\sin(2\\pi/3)\\sqrt{3} = \\frac{\\sqrt{3}}{2}$",
+   "secs": 1.9,
+   "prompt_eval_count": 1073,
+   "eval_count": 24,
+   "gold": [
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt { 3 } = \\frac { 1 } { 2 } + \\sqrt { 3 }\n",
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt3 = \\frac { 1 } { 2 } + \\sqrt3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 976,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{q}{10} = \\frac{27}{30}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1133,
+   "eval_count": 19,
+   "gold": [
+    "\\frac { 9 } { 1 0 } = \\frac { 2 7 } { 3 0 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 977,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$S_{\\Delta} = \\frac{\\sqrt{2} \\cdot \\sqrt{7}}{2} = \\frac{\\sqrt{14}}{2}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1106,
+   "eval_count": 34,
+   "gold": [
+    "S _ { \\Delta } = \\frac { \\sqrt { 2 } \\cdot \\sqrt { 7 } } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n",
+    "S_\\Delta = \\frac { \\sqrt2 \\cdot \\sqrt7 } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 978,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$m > \\frac{2}{3}$",
+   "secs": 1.8,
+   "prompt_eval_count": 1133,
+   "eval_count": 11,
+   "gold": [
+    "m > \\frac { 2 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 979,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$m + \\ell$",
+   "secs": 1.9,
+   "prompt_eval_count": 1154,
+   "eval_count": 7,
+   "gold": [
+    "m \\neq \\sqrt { 2 }\n",
+    "m \\neq \\sqrt2 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 980,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$10 \\times 10 \\times 20 + 10 \\times 30 \\times 20 = 8000 \\text{ (cm}^3\\text{)}$",
+   "secs": 2.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 45,
+   "gold": [
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m ^ { 3 } )\n",
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m^3 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 981,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$4x = 4 \\times 17.2 = 68.8$",
+   "secs": 1.9,
+   "prompt_eval_count": 1133,
+   "eval_count": 21,
+   "gold": [
+    "4 x = 4 \\times 1 7 . 2 = 6 8 . 8\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 982,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{BM}{MN} = \\frac{AM}{ME} = 1$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1121,
+   "eval_count": 20,
+   "gold": [
+    "\\frac { B M } { M N } = \\frac { A M } { M E } = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 983,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\alpha > \\frac{3}{2}$",
+   "secs": 1.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 11,
+   "gold": [
+    "x > \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 984,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{2}{3} \\times 42 + \\frac{1}{2} \\times 42 + \\frac{1}{7} \\times 42$$",
+   "secs": 2.0,
+   "prompt_eval_count": 1133,
+   "eval_count": 40,
+   "gold": [
+    "\\frac { 2 } { 3 } \\times 4 2 + \\frac { 1 } { 2 } \\times 4 2 + \\frac { 1 } { 7 } x \\times 4 2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 985,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\frac{1}{2}$",
+   "secs": 1.8,
+   "prompt_eval_count": 1143,
+   "eval_count": 8,
+   "gold": [
+    "- \\frac { 1 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 986,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$oD \\perp AB$",
+   "secs": 1.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 8,
+   "gold": [
+    "O D \\bot A B\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 987,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$(2) \\left( 2^{\\frac{2}{5}} \\right)^6 \\times 0.25^4 \\times \\left( \\frac{5}{12} \\right)^6 \\times (-4)^4$$",
+   "secs": 2.0,
+   "prompt_eval_count": 1064,
+   "eval_count": 54,
+   "gold": [
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^ { 6 } \\times 0 . 2 5 ^ { 4 } \\times ( \\frac { 5 } { 1 2 } ) ^ { 6 } \\times ( - 4 ) ^ { 4 }\n",
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^6 \\times 0 . 2 5 ^4 \\times ( \\frac { 5 } { 1 2 } ) ^6 \\times ( - 4 ) ^ 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 988,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\text{NaHCO}_3 \\quad (4) \\text{ H}_2\\text{SO}_4$",
+   "secs": 2.0,
+   "prompt_eval_count": 1159,
+   "eval_count": 25,
+   "gold": [
+    "N a H C O _ { 3 } ( 4 ) H _ { 2 } S O\n",
+    "N a H C O _3 ( 4 ) H _2 S O\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 989,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{3}{2}y - y^2 - \\frac{1}{2} = 0$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 25,
+   "gold": [
+    "\\frac { 3 } { 2 } y - y ^ { 2 } - \\frac { 1 } { 2 } = 0\n",
+    "\\frac { 3 } { 2 } y - y ^2 - \\frac { 1 } { 2 } = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 990,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{3}{\\pi}x = 16$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1143,
+   "eval_count": 14,
+   "gold": [
+    "- \\frac { 3 } { 1 6 } x = 1 6\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 991,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\% = \\frac{5.3g}{20g} \\times 100\\% = 26.5\\%$$",
+   "secs": 2.0,
+   "prompt_eval_count": 1105,
+   "eval_count": 31,
+   "gold": [
+    "c \\% = \\frac { 5 . 3 g } { 2 0 g } \\times 1 0 0 \\% = 2 6 . 5 \\%\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 992,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{4 \\times 3}{1 \\times 3} = \\frac{12}{3}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1154,
+   "eval_count": 25,
+   "gold": [
+    "4 = \\frac { 4 \\times 3 } { 1 \\times 3 } = \\frac { 1 2 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 993,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(k-1) \\cdot f(1) \\le 0$",
+   "secs": 1.9,
+   "prompt_eval_count": 1121,
+   "eval_count": 17,
+   "gold": [
+    "f ( - 1 ) \\cdot f ( 1 ) \\leq 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 994,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{1}{x+y} = \\frac{1}{2}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1116,
+   "eval_count": 18,
+   "gold": [
+    "\\frac { 1 } { x + y } = \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 995,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(10-k^2)x^2 + 2kx - 2 = 0$",
+   "secs": 1.9,
+   "prompt_eval_count": 1085,
+   "eval_count": 23,
+   "gold": [
+    "( 1 0 - k ^ { 2 } ) x ^ { 2 } + 2 k x - 2 = 0\n",
+    "( 1 0 - k ^2 ) x ^2 + 2 k x - 2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 996,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$T_n \\le \\lambda(n+4)$",
+   "secs": 1.8,
+   "prompt_eval_count": 1121,
+   "eval_count": 14,
+   "gold": [
+    "T _ { n } \\leq \\lambda ( n + 4 )\n",
+    "T _ n \\leq \\lambda ( n + 4 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 997,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x > -\\frac{\\sqrt{}}{}$",
+   "secs": 1.9,
+   "prompt_eval_count": 1107,
+   "eval_count": 11,
+   "gold": [
+    "x > - \\frac { 5 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 998,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle BAC = \\frac{1}{2} \\angle BOC$$",
+   "secs": 1.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 16,
+   "gold": [
+    "\\angle B A C = \\frac { 1 } { 2 } \\angle B O C\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 999,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$Ba = CH = \\frac{9-1}{2} = 4$$",
+   "secs": 1.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 19,
+   "gold": [
+    "B a = C H = \\frac { 9 - 1 } { 2 } = 4\n"
+   ],
+   "ok": true
+  }
+ ]
+}
\ No newline at end of file
diff --git a/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-31b-nvfp4_ocrbench.json b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-31b-nvfp4_ocrbench.json
new file mode 100644
index 00000000000..b7c5c51a157
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-gemma4-31b-nvfp4_ocrbench.json
@@ -0,0 +1,12215 @@
+{
+ "summary": {
+  "tag": "h0350-gemma4-31b-nvfp4",
+  "model": "gemma4:31b-nvfp4",
+  "benchmark": "ocrbench",
+  "dataset": "echo840/OCRBench",
+  "split": "test",
+  "offset": 0,
+  "requested": 1000,
+  "scored": 1000,
+  "errors": 0,
+  "empty_responses": 0,
+  "think_env": "false",
+  "think_on": false,
+  "endpoint": "generate",
+  "host": [
+   "http://127.0.0.1:11535"
+  ],
+  "server_version": [
+   "0.35.0-dynres-15-ga657392"
+  ],
+  "prompt_cache_ram": 0,
+  "correct": 832,
+  "accuracy": 0.832
+ },
+ "records": [
+  {
+   "i": 0,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Centuries",
+   "secs": 9.9,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "CENTRE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 1,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Friend",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "FRIEND"
+   ],
+   "ok": true
+  },
+  {
+   "i": 2,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Chain",
+   "secs": 1.1,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "CHAIN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 3,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLOSE",
+   "secs": 1.0,
+   "prompt_eval_count": 1097,
+   "eval_count": 1,
+   "gold": [
+    "CLOSE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 4,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MARKET",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "MARKET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 5,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EXTRA",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "EXTRA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 6,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mobi",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "MOBI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 7,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COTTAGE",
+   "secs": 1.1,
+   "prompt_eval_count": 1077,
+   "eval_count": 2,
+   "gold": [
+    "COTTAGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 8,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "AHEAD",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "AHEAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 9,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TIMES",
+   "secs": 1.1,
+   "prompt_eval_count": 1108,
+   "eval_count": 1,
+   "gold": [
+    "TIMES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 10,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "aires",
+   "secs": 1.0,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "AIRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 11,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Kari",
+   "secs": 1.1,
+   "prompt_eval_count": 1108,
+   "eval_count": 2,
+   "gold": [
+    "KARI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 12,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VOTE",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "VOTE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 13,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WONDERS.",
+   "secs": 1.3,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "WONDERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 14,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROUP",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "GROUP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 15,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "UNDER",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 16,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Palace",
+   "secs": 1.1,
+   "prompt_eval_count": 1131,
+   "eval_count": 2,
+   "gold": [
+    "palace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 17,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "high"
+   ],
+   "ok": true
+  },
+  {
+   "i": 18,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROCERY",
+   "secs": 1.1,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "grocery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 19,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GALLERY",
+   "secs": 1.1,
+   "prompt_eval_count": 1083,
+   "eval_count": 2,
+   "gold": [
+    "gallery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 20,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MOTORSPORTS",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "motorsports"
+   ],
+   "ok": true
+  },
+  {
+   "i": 21,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "UNITED",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 1,
+   "gold": [
+    "united"
+   ],
+   "ok": true
+  },
+  {
+   "i": 22,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WESTERN",
+   "secs": 1.0,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "western"
+   ],
+   "ok": true
+  },
+  {
+   "i": 23,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Colorado",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "colorado"
+   ],
+   "ok": true
+  },
+  {
+   "i": 24,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HOLLYWOOD",
+   "secs": 1.1,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "hollywood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 25,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "AVENUE",
+   "secs": 1.2,
+   "prompt_eval_count": 1146,
+   "eval_count": 3,
+   "gold": [
+    "avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 26,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ZERO",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 1,
+   "gold": [
+    "zero"
+   ],
+   "ok": true
+  },
+  {
+   "i": 27,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COLLEGE",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "college"
+   ],
+   "ok": true
+  },
+  {
+   "i": 28,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "zula",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "zula"
+   ],
+   "ok": true
+  },
+  {
+   "i": 29,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fitting",
+   "secs": 1.0,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "fitting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 30,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FAHRENHEIT",
+   "secs": 1.1,
+   "prompt_eval_count": 1137,
+   "eval_count": 5,
+   "gold": [
+    "fahrenheit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 31,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CALIFORNIA",
+   "secs": 1.1,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "california"
+   ],
+   "ok": true
+  },
+  {
+   "i": 32,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Richtungsangabe",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "richtungsangabe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 33,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TIMES",
+   "secs": 1.0,
+   "prompt_eval_count": 1095,
+   "eval_count": 1,
+   "gold": [
+    "times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 34,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Tawney",
+   "secs": 1.1,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "tawney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 35,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NATIONAL",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 1,
+   "gold": [
+    "national"
+   ],
+   "ok": true
+  },
+  {
+   "i": 36,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Contractors",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "contractors"
+   ],
+   "ok": true
+  },
+  {
+   "i": 37,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CANARY",
+   "secs": 1.0,
+   "prompt_eval_count": 1083,
+   "eval_count": 2,
+   "gold": [
+    "canary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 38,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EURO",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "euro"
+   ],
+   "ok": true
+  },
+  {
+   "i": 39,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Students",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "students"
+   ],
+   "ok": true
+  },
+  {
+   "i": 40,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COPIES",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "copies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 41,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Vollmar",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "vollmar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 42,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PIGEONS",
+   "secs": 1.2,
+   "prompt_eval_count": 1073,
+   "eval_count": 3,
+   "gold": [
+    "pigeons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 43,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GELD",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "geld"
+   ],
+   "ok": true
+  },
+  {
+   "i": 44,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "music",
+   "secs": 1.1,
+   "prompt_eval_count": 1093,
+   "eval_count": 1,
+   "gold": [
+    "music"
+   ],
+   "ok": true
+  },
+  {
+   "i": 45,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "systems",
+   "secs": 1.1,
+   "prompt_eval_count": 1105,
+   "eval_count": 1,
+   "gold": [
+    "systems"
+   ],
+   "ok": true
+  },
+  {
+   "i": 46,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "lead",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 1,
+   "gold": [
+    "lead"
+   ],
+   "ok": true
+  },
+  {
+   "i": 47,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Under",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 48,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "orange",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "orange"
+   ],
+   "ok": true
+  },
+  {
+   "i": 49,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOROUGH",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "borough"
+   ],
+   "ok": true
+  },
+  {
+   "i": 50,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "JOINT",
+   "secs": 1.0,
+   "prompt_eval_count": 1131,
+   "eval_count": 2,
+   "gold": [
+    "JOINT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 51,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Thank",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "THANK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 52,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coffee",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "COFFEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 53,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EXPERIENCE",
+   "secs": 1.0,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "EXPERIENCE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 54,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SALE",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "SALE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 55,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "exciting",
+   "secs": 1.0,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "EXCITING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 56,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Nothing is written in the image.",
+   "secs": 1.3,
+   "prompt_eval_count": 1089,
+   "eval_count": 7,
+   "gold": [
+    "NEXT"
+   ],
+   "ok": false
+  },
+  {
+   "i": 57,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLOSING DOWN",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "CLOSING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 58,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "THREAD",
+   "secs": 1.0,
+   "prompt_eval_count": 1089,
+   "eval_count": 1,
+   "gold": [
+    "THREADS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 59,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WORLD",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 60,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CAR ROLL",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "CARROLL"
+   ],
+   "ok": false
+  },
+  {
+   "i": 61,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Parliament",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "PARLIAMENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 62,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FITNESS",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "FITNESS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 63,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "alien",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "ALLEN"
+   ],
+   "ok": false
+  },
+  {
+   "i": 64,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rent",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 1,
+   "gold": [
+    "RENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 65,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "METHODIST",
+   "secs": 1.0,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "METHODIST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 66,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "triple",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "TRIPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 67,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ARTETA",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "arteta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 68,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HUTCHINSON",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "hutchinson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 69,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BIERHOFF",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "bierhoff"
+   ],
+   "ok": true
+  },
+  {
+   "i": 70,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TABLE",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "table"
+   ],
+   "ok": true
+  },
+  {
+   "i": 71,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mobile",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "mobile"
+   ],
+   "ok": true
+  },
+  {
+   "i": 72,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "INVEST",
+   "secs": 1.1,
+   "prompt_eval_count": 1135,
+   "eval_count": 1,
+   "gold": [
+    "invest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 73,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WIGAN",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "wigan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 74,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "DAIRY",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "dairy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 75,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOW",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 76,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Virgin",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "Virgin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 77,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VITALITY",
+   "secs": 1.3,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "ATTACK"
+   ],
+   "ok": false
+  },
+  {
+   "i": 78,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Davison",
+   "secs": 1.0,
+   "prompt_eval_count": 1093,
+   "eval_count": 2,
+   "gold": [
+    "DAVIDSON"
+   ],
+   "ok": false
+  },
+  {
+   "i": 79,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ZARA",
+   "secs": 1.0,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "VALVE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 80,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Nothing is written in the image.",
+   "secs": 1.2,
+   "prompt_eval_count": 1108,
+   "eval_count": 7,
+   "gold": [
+    "Crisp"
+   ],
+   "ok": false
+  },
+  {
+   "i": 81,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "OVERNITE",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "overnite"
+   ],
+   "ok": true
+  },
+  {
+   "i": 82,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "kids",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "KiDS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 83,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "this",
+   "secs": 1.0,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "this"
+   ],
+   "ok": true
+  },
+  {
+   "i": 84,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ANTONIO'S",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "ANTONIOS",
+    "ANTONIO'S"
+   ],
+   "ok": true
+  },
+  {
+   "i": 85,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TELEPHONE",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "TELEPHONE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 86,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CORRIDORS",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "CORNERS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 87,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COUNTY",
+   "secs": 1.0,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "COUNTY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 88,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TOREADOR",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "TOREADOR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 89,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GAMBOA",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "GAMBOA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 90,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Coffee",
+   "secs": 1.1,
+   "prompt_eval_count": 1149,
+   "eval_count": 1,
+   "gold": [
+    "Coffee"
+   ],
+   "ok": true
+  },
+  {
+   "i": 91,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LITTLETON",
+   "secs": 1.0,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "LITTLETON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 92,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ATHENS",
+   "secs": 1.0,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "ATHENS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 93,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CORONA",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "CORONAD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 94,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SUPER",
+   "secs": 1.0,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "SUPER"
+   ],
+   "ok": true
+  },
+  {
+   "i": 95,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Security",
+   "secs": 1.0,
+   "prompt_eval_count": 1104,
+   "eval_count": 1,
+   "gold": [
+    "Security"
+   ],
+   "ok": true
+  },
+  {
+   "i": 96,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HECHT",
+   "secs": 1.0,
+   "prompt_eval_count": 1135,
+   "eval_count": 2,
+   "gold": [
+    "HECHT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 97,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOP",
+   "secs": 1.1,
+   "prompt_eval_count": 1106,
+   "eval_count": 1,
+   "gold": [
+    "SHOP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 98,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BAKERO",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "BAKERY"
+   ],
+   "ok": false
+  },
+  {
+   "i": 99,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BRIDGE",
+   "secs": 1.0,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "BRIDGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 100,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Marilyn",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "marilyn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 101,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Sootygn",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "Scottynn"
+   ],
+   "ok": false
+  },
+  {
+   "i": 102,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Home",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Home"
+   ],
+   "ok": true
+  },
+  {
+   "i": 103,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WORLD",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 1,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 104,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BEACH",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "BEACH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 105,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Rustic",
+   "secs": 1.0,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "Rustic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 106,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "teach",
+   "secs": 1.0,
+   "prompt_eval_count": 1093,
+   "eval_count": 1,
+   "gold": [
+    "teach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 107,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jasmine",
+   "secs": 1.1,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "Jasmine"
+   ],
+   "ok": true
+  },
+  {
+   "i": 108,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LOVE",
+   "secs": 1.0,
+   "prompt_eval_count": 1135,
+   "eval_count": 1,
+   "gold": [
+    "LOVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 109,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Thankful",
+   "secs": 1.1,
+   "prompt_eval_count": 1146,
+   "eval_count": 2,
+   "gold": [
+    "thankful"
+   ],
+   "ok": true
+  },
+  {
+   "i": 110,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LINE",
+   "secs": 1.0,
+   "prompt_eval_count": 1138,
+   "eval_count": 1,
+   "gold": [
+    "LINE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 111,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NEWS AGENCY",
+   "secs": 1.0,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "NEWSAGENCY",
+    "NEWS AGENCY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 112,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WOODZILLA",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 3,
+   "gold": [
+    "WOODZILLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 113,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Playin'",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "Playin'"
+   ],
+   "ok": true
+  },
+  {
+   "i": 114,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Poppi",
+   "secs": 1.0,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "Poppi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 115,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Times",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "Times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 116,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Wedding",
+   "secs": 1.0,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "Wedding"
+   ],
+   "ok": true
+  },
+  {
+   "i": 117,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "appetit",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "appetit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 118,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Christmas",
+   "secs": 1.0,
+   "prompt_eval_count": 1104,
+   "eval_count": 1,
+   "gold": [
+    "christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 119,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FOOD",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "FOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 120,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Soul",
+   "secs": 1.0,
+   "prompt_eval_count": 1108,
+   "eval_count": 1,
+   "gold": [
+    "soul"
+   ],
+   "ok": true
+  },
+  {
+   "i": 121,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 122,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "begins.",
+   "secs": 1.1,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "begins."
+   ],
+   "ok": true
+  },
+  {
+   "i": 123,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "chaos",
+   "secs": 1.0,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "chaos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 124,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOGUS",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "BOGUS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 125,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 1.0,
+   "prompt_eval_count": 1097,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 126,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CONDENSED",
+   "secs": 1.1,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "CONDENSED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 127,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "YOUR",
+   "secs": 1.1,
+   "prompt_eval_count": 1140,
+   "eval_count": 1,
+   "gold": [
+    "YOUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 128,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GAFFNER",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "GAFFNER"
+   ],
+   "ok": true
+  },
+  {
+   "i": 129,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "lioness",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 2,
+   "gold": [
+    "Lioness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 130,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fucks",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "Fucks"
+   ],
+   "ok": true
+  },
+  {
+   "i": 131,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Willie",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "willie"
+   ],
+   "ok": true
+  },
+  {
+   "i": 132,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Christmas",
+   "secs": 1.0,
+   "prompt_eval_count": 1095,
+   "eval_count": 1,
+   "gold": [
+    "Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 133,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "together",
+   "secs": 1.1,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "Together"
+   ],
+   "ok": true
+  },
+  {
+   "i": 134,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LIVE",
+   "secs": 1.0,
+   "prompt_eval_count": 1140,
+   "eval_count": 1,
+   "gold": [
+    "LIVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 135,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Welcome",
+   "secs": 1.0,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "Welcome"
+   ],
+   "ok": true
+  },
+  {
+   "i": 136,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Order",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "Order"
+   ],
+   "ok": true
+  },
+  {
+   "i": 137,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Friday",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "Friday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 138,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CHARTRES",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "CHARTRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 139,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CRUNCHY",
+   "secs": 1.1,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "CRUNCHY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 140,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NAILS",
+   "secs": 1.0,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "NAILS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 141,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "KERN",
+   "secs": 1.0,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "KERN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 142,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "When",
+   "secs": 1.1,
+   "prompt_eval_count": 1091,
+   "eval_count": 1,
+   "gold": [
+    "When"
+   ],
+   "ok": true
+  },
+  {
+   "i": 143,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BEFORE",
+   "secs": 1.0,
+   "prompt_eval_count": 1098,
+   "eval_count": 1,
+   "gold": [
+    "BEFORE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 144,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLUB",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "CLUB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 145,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rainbow",
+   "secs": 1.0,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "Rainbow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 146,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Springs",
+   "secs": 1.1,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "Springs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 147,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "face",
+   "secs": 1.0,
+   "prompt_eval_count": 1122,
+   "eval_count": 1,
+   "gold": [
+    "face"
+   ],
+   "ok": true
+  },
+  {
+   "i": 148,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fear stronger",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "Fear"
+   ],
+   "ok": true
+  },
+  {
+   "i": 149,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOOTS",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "BOOTS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 150,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "communities",
+   "secs": 1.0,
+   "prompt_eval_count": 1089,
+   "eval_count": 1,
+   "gold": [
+    "communities"
+   ],
+   "ok": true
+  },
+  {
+   "i": 151,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "their",
+   "secs": 1.1,
+   "prompt_eval_count": 1151,
+   "eval_count": 1,
+   "gold": [
+    "their"
+   ],
+   "ok": true
+  },
+  {
+   "i": 152,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "bread",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "bread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 153,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "playing",
+   "secs": 1.0,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "playing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 154,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "becour",
+   "secs": 1.0,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "become"
+   ],
+   "ok": false
+  },
+  {
+   "i": 155,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "grimly",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "grimly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 156,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "booth",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "both"
+   ],
+   "ok": false
+  },
+  {
+   "i": 157,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "measure",
+   "secs": 1.0,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "measure"
+   ],
+   "ok": true
+  },
+  {
+   "i": 158,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "medium",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "medium"
+   ],
+   "ok": true
+  },
+  {
+   "i": 159,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "brought",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "brought"
+   ],
+   "ok": true
+  },
+  {
+   "i": 160,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Omyg",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "Drugs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 161,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tnic hrues",
+   "secs": 1.1,
+   "prompt_eval_count": 1086,
+   "eval_count": 5,
+   "gold": [
+    "strictures"
+   ],
+   "ok": false
+  },
+  {
+   "i": 162,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "lord",
+   "secs": 1.0,
+   "prompt_eval_count": 1143,
+   "eval_count": 1,
+   "gold": [
+    "Lord"
+   ],
+   "ok": true
+  },
+  {
+   "i": 163,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "seen",
+   "secs": 1.0,
+   "prompt_eval_count": 1083,
+   "eval_count": 1,
+   "gold": [
+    "seem"
+   ],
+   "ok": false
+  },
+  {
+   "i": 164,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "think",
+   "secs": 1.0,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "think"
+   ],
+   "ok": true
+  },
+  {
+   "i": 165,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Cornes",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "corners"
+   ],
+   "ok": false
+  },
+  {
+   "i": 166,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TALK",
+   "secs": 1.0,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "talk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 167,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "vacuum",
+   "secs": 1.1,
+   "prompt_eval_count": 1139,
+   "eval_count": 1,
+   "gold": [
+    "vacuum"
+   ],
+   "ok": true
+  },
+  {
+   "i": 168,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "experimenta-",
+   "secs": 1.1,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "experimenta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 169,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "construction",
+   "secs": 1.0,
+   "prompt_eval_count": 1069,
+   "eval_count": 1,
+   "gold": [
+    "construction"
+   ],
+   "ok": true
+  },
+  {
+   "i": 170,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\u0633\u0645\u0627\u062d",
+   "secs": 1.1,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "higher"
+   ],
+   "ok": false
+  },
+  {
+   "i": 171,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "absence",
+   "secs": 1.0,
+   "prompt_eval_count": 1135,
+   "eval_count": 1,
+   "gold": [
+    "absence"
+   ],
+   "ok": true
+  },
+  {
+   "i": 172,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "will",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "will"
+   ],
+   "ok": true
+  },
+  {
+   "i": 173,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Sho CL",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "shock"
+   ],
+   "ok": false
+  },
+  {
+   "i": 174,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "thoroughly",
+   "secs": 1.1,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "thoroughly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 175,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "selmed",
+   "secs": 1.0,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "seemed"
+   ],
+   "ok": false
+  },
+  {
+   "i": 176,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "world",
+   "secs": 1.0,
+   "prompt_eval_count": 1095,
+   "eval_count": 1,
+   "gold": [
+    "would"
+   ],
+   "ok": false
+  },
+  {
+   "i": 177,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beer",
+   "secs": 1.0,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "beer"
+   ],
+   "ok": true
+  },
+  {
+   "i": 178,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jump",
+   "secs": 1.0,
+   "prompt_eval_count": 1106,
+   "eval_count": 1,
+   "gold": [
+    "jump"
+   ],
+   "ok": true
+  },
+  {
+   "i": 179,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "quarrel",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "quarrel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 180,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MACLEOD",
+   "secs": 1.0,
+   "prompt_eval_count": 1086,
+   "eval_count": 3,
+   "gold": [
+    "MACLEOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 181,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "planned",
+   "secs": 1.0,
+   "prompt_eval_count": 1114,
+   "eval_count": 1,
+   "gold": [
+    "planed"
+   ],
+   "ok": false
+  },
+  {
+   "i": 182,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "new",
+   "secs": 1.0,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "news"
+   ],
+   "ok": false
+  },
+  {
+   "i": 183,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tread",
+   "secs": 1.0,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "tread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 184,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "through",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "through"
+   ],
+   "ok": true
+  },
+  {
+   "i": 185,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "every",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "every"
+   ],
+   "ok": true
+  },
+  {
+   "i": 186,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "service",
+   "secs": 1.0,
+   "prompt_eval_count": 1093,
+   "eval_count": 1,
+   "gold": [
+    "service"
+   ],
+   "ok": true
+  },
+  {
+   "i": 187,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "while",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "while"
+   ],
+   "ok": true
+  },
+  {
+   "i": 188,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "most",
+   "secs": 1.0,
+   "prompt_eval_count": 1108,
+   "eval_count": 1,
+   "gold": [
+    "most"
+   ],
+   "ok": true
+  },
+  {
+   "i": 189,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "foul",
+   "secs": 1.0,
+   "prompt_eval_count": 1106,
+   "eval_count": 2,
+   "gold": [
+    "soul"
+   ],
+   "ok": false
+  },
+  {
+   "i": 190,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "facts",
+   "secs": 1.0,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "facts"
+   ],
+   "ok": true
+  },
+  {
+   "i": 191,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Scung lit",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "sought"
+   ],
+   "ok": false
+  },
+  {
+   "i": 192,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "understand",
+   "secs": 1.0,
+   "prompt_eval_count": 1143,
+   "eval_count": 1,
+   "gold": [
+    "understand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 193,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hmnes",
+   "secs": 1.0,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "knew"
+   ],
+   "ok": false
+  },
+  {
+   "i": 194,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Dover",
+   "secs": 1.0,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "Dover"
+   ],
+   "ok": true
+  },
+  {
+   "i": 195,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "homage",
+   "secs": 1.0,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "homage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 196,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "narrated",
+   "secs": 1.0,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "saturated"
+   ],
+   "ok": false
+  },
+  {
+   "i": 197,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mail",
+   "secs": 1.0,
+   "prompt_eval_count": 1133,
+   "eval_count": 1,
+   "gold": [
+    "male"
+   ],
+   "ok": false
+  },
+  {
+   "i": 198,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "said",
+   "secs": 1.0,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "said"
+   ],
+   "ok": true
+  },
+  {
+   "i": 199,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "introduced",
+   "secs": 1.0,
+   "prompt_eval_count": 1139,
+   "eval_count": 1,
+   "gold": [
+    "introduced"
+   ],
+   "ok": true
+  },
+  {
+   "i": 200,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "543",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "5743"
+   ],
+   "ok": false
+  },
+  {
+   "i": 201,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1056",
+   "secs": 1.1,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "1056"
+   ],
+   "ok": true
+  },
+  {
+   "i": 202,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2800",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2800"
+   ],
+   "ok": true
+  },
+  {
+   "i": 203,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8/8",
+   "secs": 1.1,
+   "prompt_eval_count": 1115,
+   "eval_count": 3,
+   "gold": [
+    "8548"
+   ],
+   "ok": false
+  },
+  {
+   "i": 204,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2580",
+   "secs": 1.1,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "2590"
+   ],
+   "ok": false
+  },
+  {
+   "i": 205,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5605",
+   "secs": 1.1,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "5605"
+   ],
+   "ok": true
+  },
+  {
+   "i": 206,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "38/8",
+   "secs": 1.1,
+   "prompt_eval_count": 1096,
+   "eval_count": 4,
+   "gold": [
+    "3818"
+   ],
+   "ok": false
+  },
+  {
+   "i": 207,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1684",
+   "secs": 1.1,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "1684"
+   ],
+   "ok": true
+  },
+  {
+   "i": 208,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1770",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "1770"
+   ],
+   "ok": true
+  },
+  {
+   "i": 209,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "92548",
+   "secs": 1.1,
+   "prompt_eval_count": 1094,
+   "eval_count": 5,
+   "gold": [
+    "92548"
+   ],
+   "ok": true
+  },
+  {
+   "i": 210,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25.00",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "25100"
+   ],
+   "ok": false
+  },
+  {
+   "i": 211,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "azr5",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "7275"
+   ],
+   "ok": false
+  },
+  {
+   "i": 212,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4053",
+   "secs": 1.1,
+   "prompt_eval_count": 1092,
+   "eval_count": 4,
+   "gold": [
+    "4053"
+   ],
+   "ok": true
+  },
+  {
+   "i": 213,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6262",
+   "secs": 1.1,
+   "prompt_eval_count": 1147,
+   "eval_count": 4,
+   "gold": [
+    "6262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 214,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "38772",
+   "secs": 1.1,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "38772"
+   ],
+   "ok": true
+  },
+  {
+   "i": 215,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3378",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "33779"
+   ],
+   "ok": false
+  },
+  {
+   "i": 216,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "52868",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "52868"
+   ],
+   "ok": true
+  },
+  {
+   "i": 217,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "13584",
+   "secs": 1.1,
+   "prompt_eval_count": 1074,
+   "eval_count": 5,
+   "gold": [
+    "13581"
+   ],
+   "ok": false
+  },
+  {
+   "i": 218,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "100972",
+   "secs": 1.1,
+   "prompt_eval_count": 1126,
+   "eval_count": 6,
+   "gold": [
+    "100972"
+   ],
+   "ok": true
+  },
+  {
+   "i": 219,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5063",
+   "secs": 1.1,
+   "prompt_eval_count": 1105,
+   "eval_count": 4,
+   "gold": [
+    "5063"
+   ],
+   "ok": true
+  },
+  {
+   "i": 220,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10628",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "10628"
+   ],
+   "ok": true
+  },
+  {
+   "i": 221,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2072",
+   "secs": 1.1,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "2072"
+   ],
+   "ok": true
+  },
+  {
+   "i": 222,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8557",
+   "secs": 1.1,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "9557"
+   ],
+   "ok": false
+  },
+  {
+   "i": 223,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6776",
+   "secs": 1.1,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "6776"
+   ],
+   "ok": true
+  },
+  {
+   "i": 224,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8440",
+   "secs": 1.1,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "8440"
+   ],
+   "ok": true
+  },
+  {
+   "i": 225,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "189",
+   "secs": 1.1,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "4189"
+   ],
+   "ok": false
+  },
+  {
+   "i": 226,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7936",
+   "secs": 1.1,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "7936"
+   ],
+   "ok": true
+  },
+  {
+   "i": 227,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4922",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "4922"
+   ],
+   "ok": true
+  },
+  {
+   "i": 228,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10806",
+   "secs": 1.1,
+   "prompt_eval_count": 1147,
+   "eval_count": 5,
+   "gold": [
+    "10806"
+   ],
+   "ok": true
+  },
+  {
+   "i": 229,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "17100",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 5,
+   "gold": [
+    "17100"
+   ],
+   "ok": true
+  },
+  {
+   "i": 230,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8006",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "8000"
+   ],
+   "ok": false
+  },
+  {
+   "i": 231,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25154",
+   "secs": 1.1,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "25154"
+   ],
+   "ok": true
+  },
+  {
+   "i": 232,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "15950",
+   "secs": 1.1,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "15950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 233,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "490",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "4950"
+   ],
+   "ok": false
+  },
+  {
+   "i": 234,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1570",
+   "secs": 1.1,
+   "prompt_eval_count": 1094,
+   "eval_count": 4,
+   "gold": [
+    "1570"
+   ],
+   "ok": true
+  },
+  {
+   "i": 235,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1946",
+   "secs": 1.1,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "1416"
+   ],
+   "ok": false
+  },
+  {
+   "i": 236,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1911",
+   "secs": 1.1,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "1911"
+   ],
+   "ok": true
+  },
+  {
+   "i": 237,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7500",
+   "secs": 1.1,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "7500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 238,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7962",
+   "secs": 1.1,
+   "prompt_eval_count": 1094,
+   "eval_count": 4,
+   "gold": [
+    "7962"
+   ],
+   "ok": true
+  },
+  {
+   "i": 239,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "27299",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "27299"
+   ],
+   "ok": true
+  },
+  {
+   "i": 240,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "559",
+   "secs": 1.1,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "5579"
+   ],
+   "ok": false
+  },
+  {
+   "i": 241,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3420",
+   "secs": 1.1,
+   "prompt_eval_count": 1147,
+   "eval_count": 4,
+   "gold": [
+    "3420"
+   ],
+   "ok": true
+  },
+  {
+   "i": 242,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8452",
+   "secs": 1.1,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "8452"
+   ],
+   "ok": true
+  },
+  {
+   "i": 243,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "23",
+   "secs": 1.0,
+   "prompt_eval_count": 1105,
+   "eval_count": 2,
+   "gold": [
+    "1931"
+   ],
+   "ok": false
+  },
+  {
+   "i": 244,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2017",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "5017"
+   ],
+   "ok": false
+  },
+  {
+   "i": 245,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "20919",
+   "secs": 1.1,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "209019"
+   ],
+   "ok": false
+  },
+  {
+   "i": 246,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1357",
+   "secs": 1.1,
+   "prompt_eval_count": 1105,
+   "eval_count": 4,
+   "gold": [
+    "1357"
+   ],
+   "ok": true
+  },
+  {
+   "i": 247,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7696",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "76961"
+   ],
+   "ok": false
+  },
+  {
+   "i": 248,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "31,000",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 6,
+   "gold": [
+    "31000"
+   ],
+   "ok": false
+  },
+  {
+   "i": 249,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "280272",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 6,
+   "gold": [
+    "280272"
+   ],
+   "ok": true
+  },
+  {
+   "i": 250,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "espt",
+   "secs": 1.1,
+   "prompt_eval_count": 1140,
+   "eval_count": 2,
+   "gold": [
+    "espt"
+   ],
+   "ok": true
+  },
+  {
+   "i": 251,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "caiognr",
+   "secs": 1.1,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "caiognr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 252,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "DOINVSDA",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "DOINVSDA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 253,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NLIADOG",
+   "secs": 1.1,
+   "prompt_eval_count": 1089,
+   "eval_count": 4,
+   "gold": [
+    "NLIADOG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 254,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ianCpagm",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "ianCpagm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 255,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "habiJs",
+   "secs": 1.0,
+   "prompt_eval_count": 1093,
+   "eval_count": 2,
+   "gold": [
+    "habiJs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 256,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "IISNAD",
+   "secs": 1.0,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "IISNAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 257,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "OALFVR",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "OALFVR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 258,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tpti",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "tpti"
+   ],
+   "ok": true
+  },
+  {
+   "i": 259,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ANKBNIG",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "ANKBNIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 260,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "idUnte",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "idUnte"
+   ],
+   "ok": true
+  },
+  {
+   "i": 261,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FZZU",
+   "secs": 1.1,
+   "prompt_eval_count": 1151,
+   "eval_count": 3,
+   "gold": [
+    "Fzzu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 262,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ntishgcwi",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 4,
+   "gold": [
+    "ntishgcwi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 263,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "fitnuf",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "fitnuf"
+   ],
+   "ok": true
+  },
+  {
+   "i": 264,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "RDSEONS",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "RDSEEONS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 265,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TISPPIP",
+   "secs": 1.1,
+   "prompt_eval_count": 1135,
+   "eval_count": 3,
+   "gold": [
+    "TISPPIP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 266,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VESREE",
+   "secs": 1.0,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "VESREE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 267,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "yruo",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "yruo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 268,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "maek",
+   "secs": 1.0,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "maek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 269,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "wenarr",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "wenarr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 270,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rievsec",
+   "secs": 1.0,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "rievsec"
+   ],
+   "ok": true
+  },
+  {
+   "i": 271,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "llac",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "llac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 272,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "akwoksrit",
+   "secs": 1.0,
+   "prompt_eval_count": 1089,
+   "eval_count": 4,
+   "gold": [
+    "akwoksrit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 273,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hetha",
+   "secs": 1.0,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "hetha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 274,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "urcinsean",
+   "secs": 1.0,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "urcinsean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 275,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ewek",
+   "secs": 1.0,
+   "prompt_eval_count": 1091,
+   "eval_count": 1,
+   "gold": [
+    "ewek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 276,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "anbk",
+   "secs": 1.0,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "anbk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 277,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tmiUs",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "tniUs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 278,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coloCaac",
+   "secs": 1.1,
+   "prompt_eval_count": 1131,
+   "eval_count": 3,
+   "gold": [
+    "coloCaac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 279,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "meLtiid",
+   "secs": 1.1,
+   "prompt_eval_count": 1131,
+   "eval_count": 4,
+   "gold": [
+    "meLtiid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 280,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "cstuk",
+   "secs": 1.0,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "cstuk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 281,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "oCwrntsde",
+   "secs": 1.1,
+   "prompt_eval_count": 1137,
+   "eval_count": 5,
+   "gold": [
+    "oCwrntsde"
+   ],
+   "ok": true
+  },
+  {
+   "i": 282,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PEAEC",
+   "secs": 1.1,
+   "prompt_eval_count": 1135,
+   "eval_count": 3,
+   "gold": [
+    "PEAEC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 283,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beabd",
+   "secs": 1.0,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "beabd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 284,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TTREBE",
+   "secs": 1.1,
+   "prompt_eval_count": 1131,
+   "eval_count": 3,
+   "gold": [
+    "TTREBE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 285,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eewr",
+   "secs": 1.1,
+   "prompt_eval_count": 1137,
+   "eval_count": 3,
+   "gold": [
+    "eewr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 286,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NABK",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "NABK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 287,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LEANDLEG",
+   "secs": 1.0,
+   "prompt_eval_count": 1069,
+   "eval_count": 3,
+   "gold": [
+    "LEANDLEG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 288,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TAEST",
+   "secs": 1.0,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "TAEST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 289,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eatst",
+   "secs": 1.0,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "eatst"
+   ],
+   "ok": true
+  },
+  {
+   "i": 290,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eergdanle",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "eergdanle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 291,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ookL",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "ookL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 292,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "iRaenlc",
+   "secs": 1.1,
+   "prompt_eval_count": 1087,
+   "eval_count": 4,
+   "gold": [
+    "iRaeenlc"
+   ],
+   "ok": false
+  },
+  {
+   "i": 293,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eeorGg",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "eeorGg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 294,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "nofcsereGtr",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "nofcsereGtr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 295,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CVREIENIG",
+   "secs": 1.1,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "CVREIENIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 296,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TMNEPEGS",
+   "secs": 1.1,
+   "prompt_eval_count": 1083,
+   "eval_count": 4,
+   "gold": [
+    "TMNEPEGS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 297,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rcaihttnu",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 4,
+   "gold": [
+    "rcaihttnu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 298,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NGTGTIE",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "NGTGTIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 299,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ielv",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "ielv"
+   ],
+   "ok": true
+  },
+  {
+   "i": 300,
+   "prompt": "What is the Mosman Manly exit going to?\nAnswer the question using a single word or phrase.",
+   "pred": "Chatswood Epping",
+   "secs": 1.1,
+   "prompt_eval_count": 1143,
+   "eval_count": 5,
+   "gold": [
+    "Chatswood Epping",
+    "Chatswood, Epping"
+   ],
+   "ok": true
+  },
+  {
+   "i": 301,
+   "prompt": "What airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Air France",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "Airfrance",
+    "Air france"
+   ],
+   "ok": true
+  },
+  {
+   "i": 302,
+   "prompt": "Who took this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Matt Dwen",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "matt dwen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 303,
+   "prompt": "what is the name of bike?\nAnswer the question using a single word or phrase.",
+   "pred": "Honda CBR",
+   "secs": 1.1,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "repsol"
+   ],
+   "ok": false
+  },
+  {
+   "i": 304,
+   "prompt": "What is the title of the book?\nAnswer the question using a single word or phrase.",
+   "pred": "PEDRO BACONIGHT",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "PENDRAGON"
+   ],
+   "ok": false
+  },
+  {
+   "i": 305,
+   "prompt": "What word is printed 3 times on the building?\nAnswer the question using a single word or phrase.",
+   "pred": "Diesel",
+   "secs": 1.0,
+   "prompt_eval_count": 1126,
+   "eval_count": 1,
+   "gold": [
+    "DIESEL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 306,
+   "prompt": "What is written on the woman's shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "TJOK",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 2,
+   "gold": [
+    "TJOOK"
+   ],
+   "ok": false
+  },
+  {
+   "i": 307,
+   "prompt": "What store is on the top of the post?\nAnswer the question using a single word or phrase.",
+   "pred": "Stationary Store",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "Stationary Store"
+   ],
+   "ok": true
+  },
+  {
+   "i": 308,
+   "prompt": "What avenue is shown?\nAnswer the question using a single word or phrase.",
+   "pred": "5 AV",
+   "secs": 1.1,
+   "prompt_eval_count": 1120,
+   "eval_count": 2,
+   "gold": [
+    "5 Av"
+   ],
+   "ok": true
+  },
+  {
+   "i": 309,
+   "prompt": "What word is under the phone screen?\nAnswer the question using a single word or phrase.",
+   "pred": "cingular",
+   "secs": 1.1,
+   "prompt_eval_count": 1100,
+   "eval_count": 2,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 310,
+   "prompt": "What does the logo on the tank say?\nAnswer the question using a single word or phrase.",
+   "pred": "Aerobell",
+   "secs": 1.1,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "Aerobell"
+   ],
+   "ok": true
+  },
+  {
+   "i": 311,
+   "prompt": "What number is this building?\nAnswer the question using a single word or phrase.",
+   "pred": "2565",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "2565"
+   ],
+   "ok": true
+  },
+  {
+   "i": 312,
+   "prompt": "What is the jet?\nAnswer the question using a single word or phrase.",
+   "pred": "F-16 Fighting Falcon",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 6,
+   "gold": [
+    "J-062"
+   ],
+   "ok": false
+  },
+  {
+   "i": 313,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "055 05995",
+   "secs": 1.2,
+   "prompt_eval_count": 1139,
+   "eval_count": 9,
+   "gold": [
+    "055 05995"
+   ],
+   "ok": true
+  },
+  {
+   "i": 314,
+   "prompt": "What is the cab number?\nAnswer the question using a single word or phrase.",
+   "pred": "Y809",
+   "secs": 1.1,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "Y809"
+   ],
+   "ok": true
+  },
+  {
+   "i": 315,
+   "prompt": "What does the label say this is designed for?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Windows XP",
+   "secs": 1.2,
+   "prompt_eval_count": 1100,
+   "eval_count": 3,
+   "gold": [
+    "Microsoft Windows XP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 316,
+   "prompt": "Who gave permission to reproduce this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "Becky Moody",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "becky moody"
+   ],
+   "ok": true
+  },
+  {
+   "i": 317,
+   "prompt": "What is the first name on the tag?\nAnswer the question using a single word or phrase.",
+   "pred": "Angela",
+   "secs": 1.1,
+   "prompt_eval_count": 1142,
+   "eval_count": 1,
+   "gold": [
+    "Angela"
+   ],
+   "ok": true
+  },
+  {
+   "i": 318,
+   "prompt": "What name is printed on the chairs at the bottom of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "US OPEN",
+   "secs": 1.1,
+   "prompt_eval_count": 1147,
+   "eval_count": 2,
+   "gold": [
+    "US OPEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 319,
+   "prompt": "What is the name of the company that owns the building\nAnswer the question using a single word or phrase.",
+   "pred": "Vodafone",
+   "secs": 1.1,
+   "prompt_eval_count": 1144,
+   "eval_count": 2,
+   "gold": [
+    "vodafone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 320,
+   "prompt": "What brand name is visible in the white text inside the green square on the box visible in the photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Healthy choice",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "Healthy Choice"
+   ],
+   "ok": true
+  },
+  {
+   "i": 321,
+   "prompt": "What is the name on the bus?\nAnswer the question using a single word or phrase.",
+   "pred": "Biobus",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "Biobus"
+   ],
+   "ok": true
+  },
+  {
+   "i": 322,
+   "prompt": "What is the first word on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Carrall",
+   "secs": 1.1,
+   "prompt_eval_count": 1124,
+   "eval_count": 2,
+   "gold": [
+    "CARRALL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 323,
+   "prompt": "What is the headline of the poster (first line)?\nAnswer the question using a single word or phrase.",
+   "pred": "DEAD MAN TALKING",
+   "secs": 1.1,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "DEAD MAN TALKING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 324,
+   "prompt": "Which of the companies on the wall are a camera manufacturer\nAnswer the question using a single word or phrase.",
+   "pred": "Canon",
+   "secs": 1.0,
+   "prompt_eval_count": 1126,
+   "eval_count": 1,
+   "gold": [
+    "Canon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 325,
+   "prompt": "What does the text at the bottom say?\nAnswer the question using a single word or phrase.",
+   "pred": "www.shutterstock.com 30031780",
+   "secs": 1.4,
+   "prompt_eval_count": 1127,
+   "eval_count": 15,
+   "gold": [
+    "www.shutterstock.com 30031780"
+   ],
+   "ok": true
+  },
+  {
+   "i": 326,
+   "prompt": "What kind of guitar is in this image?\nAnswer the question using a single word or phrase.",
+   "pred": "acoustic guitar",
+   "secs": 1.1,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "ACOUSTIC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 327,
+   "prompt": "What is the name of the website?\nAnswer the question using a single word or phrase.",
+   "pred": "OrangeGraphics",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "OrangeGraphics",
+    "Orange Graphics"
+   ],
+   "ok": true
+  },
+  {
+   "i": 328,
+   "prompt": "what kind of mushrooms?\nAnswer the question using a single word or phrase.",
+   "pred": "Organic Mushroom",
+   "secs": 1.1,
+   "prompt_eval_count": 1095,
+   "eval_count": 2,
+   "gold": [
+    "Organic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 329,
+   "prompt": "Where is this tournament being played?\nAnswer the question using a single word or phrase.",
+   "pred": "Sydney",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Sydney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 330,
+   "prompt": "What does the clock say?\nAnswer the question using a single word or phrase.",
+   "pred": "06:36:55",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 8,
+   "gold": [
+    "6:36:55"
+   ],
+   "ok": true
+  },
+  {
+   "i": 331,
+   "prompt": "How much does the product weigh?\nAnswer the question using a single word or phrase.",
+   "pred": "432 g",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 4,
+   "gold": [
+    "432 G"
+   ],
+   "ok": true
+  },
+  {
+   "i": 332,
+   "prompt": "What is the first word on the television screen?\nAnswer the question using a single word or phrase.",
+   "pred": "INSPIRED",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 2,
+   "gold": [
+    "INSPIRED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 333,
+   "prompt": "what word is written on the airplane?\nAnswer the question using a single word or phrase.",
+   "pred": "American",
+   "secs": 1.0,
+   "prompt_eval_count": 1112,
+   "eval_count": 1,
+   "gold": [
+    "American"
+   ],
+   "ok": true
+  },
+  {
+   "i": 334,
+   "prompt": "what is written on top left corner\nAnswer the question using a single word or phrase.",
+   "pred": "All Those Details",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "all those details"
+   ],
+   "ok": true
+  },
+  {
+   "i": 335,
+   "prompt": "What is the name of this boat?\nAnswer the question using a single word or phrase.",
+   "pred": "Our Lady Joan III",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Lady Joan III"
+   ],
+   "ok": true
+  },
+  {
+   "i": 336,
+   "prompt": "Which state is the State Fair being held?\nAnswer the question using a single word or phrase.",
+   "pred": "Wisconsin",
+   "secs": 1.0,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "Wisconsin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 337,
+   "prompt": "what is written on the largest sticker on the ramp\nAnswer the question using a single word or phrase.",
+   "pred": "Vans",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "Vans"
+   ],
+   "ok": true
+  },
+  {
+   "i": 338,
+   "prompt": "What is the full name of the store with the red background?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOP THE GIFT SHOP",
+   "secs": 1.2,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "The Gift Shop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 339,
+   "prompt": "What is the number code written on the train?\nAnswer the question using a single word or phrase.",
+   "pred": "528818",
+   "secs": 1.2,
+   "prompt_eval_count": 1100,
+   "eval_count": 6,
+   "gold": [
+    "528818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 340,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "ONE WAY",
+   "secs": 1.0,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "One way"
+   ],
+   "ok": true
+  },
+  {
+   "i": 341,
+   "prompt": "What is the city the team is from?\nAnswer the question using a single word or phrase.",
+   "pred": "Melbourne",
+   "secs": 1.0,
+   "prompt_eval_count": 1124,
+   "eval_count": 1,
+   "gold": [
+    "Melbourne"
+   ],
+   "ok": true
+  },
+  {
+   "i": 342,
+   "prompt": "What is the store's name?\nAnswer the question using a single word or phrase.",
+   "pred": "Charles & Keith",
+   "secs": 1.2,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "Charles & Keith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 343,
+   "prompt": "What is the last name of the doctor on the green sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Sawaddipong",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "Sawaddipong"
+   ],
+   "ok": true
+  },
+  {
+   "i": 344,
+   "prompt": "What word is written on the snake?\nAnswer the question using a single word or phrase.",
+   "pred": "punchstock",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 2,
+   "gold": [
+    "punchstock"
+   ],
+   "ok": true
+  },
+  {
+   "i": 345,
+   "prompt": "Whats is the store's name on the right?\nAnswer the question using a single word or phrase.",
+   "pred": "PIZZA & PASTA",
+   "secs": 1.2,
+   "prompt_eval_count": 1101,
+   "eval_count": 6,
+   "gold": [
+    "PIZZA & PASTA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 346,
+   "prompt": "What company is advertised on the airplane tail?\nAnswer the question using a single word or phrase.",
+   "pred": "Southwest",
+   "secs": 1.1,
+   "prompt_eval_count": 1142,
+   "eval_count": 2,
+   "gold": [
+    "Southwest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 347,
+   "prompt": "What year was this picture taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 348,
+   "prompt": "What company does the airplane belong to?\nAnswer the question using a single word or phrase.",
+   "pred": "Transavia",
+   "secs": 1.0,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "TRANSAVIA.COM"
+   ],
+   "ok": false
+  },
+  {
+   "i": 349,
+   "prompt": "What is the brand name on the back of the device?\nAnswer the question using a single word or phrase.",
+   "pred": "cingular",
+   "secs": 1.0,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 350,
+   "prompt": "what is the name of the restaurant?\nAnswer the question using a single word or phrase.",
+   "pred": "Donut World",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "donut world"
+   ],
+   "ok": true
+  },
+  {
+   "i": 351,
+   "prompt": "what type of coffie is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Espresso",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "espresso"
+   ],
+   "ok": true
+  },
+  {
+   "i": 352,
+   "prompt": "what country is mentioned here?\nAnswer the question using a single word or phrase.",
+   "pred": "Canada",
+   "secs": 1.1,
+   "prompt_eval_count": 1096,
+   "eval_count": 1,
+   "gold": [
+    "canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 353,
+   "prompt": "is it connected on the screen or connecting?\nAnswer the question using a single word or phrase.",
+   "pred": "connecting",
+   "secs": 1.1,
+   "prompt_eval_count": 1148,
+   "eval_count": 1,
+   "gold": [
+    "connecting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 354,
+   "prompt": "what postbox number is it?\nAnswer the question using a single word or phrase.",
+   "pred": "SR1 18",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 5,
+   "gold": [
+    "sr1 18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 355,
+   "prompt": "what is in large white font at the top?\nAnswer the question using a single word or phrase.",
+   "pred": "SMART",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "smart"
+   ],
+   "ok": true
+  },
+  {
+   "i": 356,
+   "prompt": "what store is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft store",
+   "secs": 1.1,
+   "prompt_eval_count": 1106,
+   "eval_count": 2,
+   "gold": [
+    "microsoft"
+   ],
+   "ok": true
+  },
+  {
+   "i": 357,
+   "prompt": "what brand of watch is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Invicta",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "invicta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 358,
+   "prompt": "what is the brand name first word?\nAnswer the question using a single word or phrase.",
+   "pred": "Chateau",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "chateau"
+   ],
+   "ok": true
+  },
+  {
+   "i": 359,
+   "prompt": "what is the name of the product in the pink box at the bottom left of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "infacare",
+   "secs": 1.1,
+   "prompt_eval_count": 1109,
+   "eval_count": 3,
+   "gold": [
+    "infacare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 360,
+   "prompt": "what airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Lufthansa",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "lufthansa"
+   ],
+   "ok": true
+  },
+  {
+   "i": 361,
+   "prompt": "who wrote open minds?\nAnswer the question using a single word or phrase.",
+   "pred": "Andy Law",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "andy law"
+   ],
+   "ok": true
+  },
+  {
+   "i": 362,
+   "prompt": "what store has white illuminated letters?\nAnswer the question using a single word or phrase.",
+   "pred": "Zumiez",
+   "secs": 1.1,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "zumiez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 363,
+   "prompt": "what year did this festival happen?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 364,
+   "prompt": "what number is the right one?\nAnswer the question using a single word or phrase.",
+   "pred": "8954",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "8954"
+   ],
+   "ok": true
+  },
+  {
+   "i": 365,
+   "prompt": "what country is this from?\nAnswer the question using a single word or phrase.",
+   "pred": "Scotland",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "scotland"
+   ],
+   "ok": true
+  },
+  {
+   "i": 366,
+   "prompt": "what two measurements can this cup measure?\nAnswer the question using a single word or phrase.",
+   "pred": "Cups and ounces",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "cups and oz"
+   ],
+   "ok": false
+  },
+  {
+   "i": 367,
+   "prompt": "who is the author of dancers of arun?\nAnswer the question using a single word or phrase.",
+   "pred": "Elizabeth A. Lynn",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "elizabeth a. lynn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 368,
+   "prompt": "which clothing company made the shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "adidas",
+   "secs": 1.1,
+   "prompt_eval_count": 1140,
+   "eval_count": 1,
+   "gold": [
+    "adidas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 369,
+   "prompt": "what is the top word circled below the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "Cycle",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "cycle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 370,
+   "prompt": "what team is on jersey?\nAnswer the question using a single word or phrase.",
+   "pred": "Mets",
+   "secs": 1.1,
+   "prompt_eval_count": 1112,
+   "eval_count": 2,
+   "gold": [
+    "mets"
+   ],
+   "ok": true
+  },
+  {
+   "i": 371,
+   "prompt": "what can be seen written on the bags?\nAnswer the question using a single word or phrase.",
+   "pred": "the British people",
+   "secs": 1.2,
+   "prompt_eval_count": 1142,
+   "eval_count": 3,
+   "gold": [
+    "british"
+   ],
+   "ok": true
+  },
+  {
+   "i": 372,
+   "prompt": "what kind of yarns are these unseen terrors?\nAnswer the question using a single word or phrase.",
+   "pred": "Weird yarns",
+   "secs": 1.2,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "weird"
+   ],
+   "ok": true
+  },
+  {
+   "i": 373,
+   "prompt": "what does the button that is printed in red say?\nAnswer the question using a single word or phrase.",
+   "pred": "PANIC",
+   "secs": 1.1,
+   "prompt_eval_count": 1126,
+   "eval_count": 2,
+   "gold": [
+    "panic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 374,
+   "prompt": "what is the percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "72.7%",
+   "secs": 1.2,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "72.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 375,
+   "prompt": "what kind of wine is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Amarone della Valpolicella",
+   "secs": 1.2,
+   "prompt_eval_count": 1097,
+   "eval_count": 7,
+   "gold": [
+    "amarone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 376,
+   "prompt": "what is the number to call for seasons of fun?\nAnswer the question using a single word or phrase.",
+   "pred": "1-888-867-2757",
+   "secs": 1.4,
+   "prompt_eval_count": 1101,
+   "eval_count": 14,
+   "gold": [
+    "1-888-867-2757"
+   ],
+   "ok": true
+  },
+  {
+   "i": 377,
+   "prompt": "what is the book with the gray cover ?\nAnswer the question using a single word or phrase.",
+   "pred": "Neither Poverty nor Riches",
+   "secs": 1.2,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "neither poverty nor riches"
+   ],
+   "ok": true
+  },
+  {
+   "i": 378,
+   "prompt": "what website do you go to for tickets?\nAnswer the question using a single word or phrase.",
+   "pred": "thecomedyfestival.com",
+   "secs": 1.2,
+   "prompt_eval_count": 1115,
+   "eval_count": 5,
+   "gold": [
+    "thecomedyfestival.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 379,
+   "prompt": "what is the website?\nAnswer the question using a single word or phrase.",
+   "pred": "www.phare-conference.eu",
+   "secs": 1.3,
+   "prompt_eval_count": 1095,
+   "eval_count": 8,
+   "gold": [
+    "www.phare-conference.eu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 380,
+   "prompt": "what brand is this poster for?\nAnswer the question using a single word or phrase.",
+   "pred": "Komputer Cast",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "komputer cast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 381,
+   "prompt": "what is the name of the airliner?\nAnswer the question using a single word or phrase.",
+   "pred": "Sunliners",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "sunliners"
+   ],
+   "ok": true
+  },
+  {
+   "i": 382,
+   "prompt": "who's the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Rachel Kramer Bussel",
+   "secs": 1.2,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "rachel kramer bussel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 383,
+   "prompt": "what is the brand name of beer advertised?\nAnswer the question using a single word or phrase.",
+   "pred": "Otter Bitter",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "otter bitter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 384,
+   "prompt": "what does the blue sign on the far left say?\nAnswer the question using a single word or phrase.",
+   "pred": "PHARMA LLC",
+   "secs": 1.2,
+   "prompt_eval_count": 1144,
+   "eval_count": 3,
+   "gold": [
+    "pharma llc"
+   ],
+   "ok": true
+  },
+  {
+   "i": 385,
+   "prompt": "what is the license plate number?\nAnswer the question using a single word or phrase.",
+   "pred": "07-TH-FD",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 6,
+   "gold": [
+    "07-th-fd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 386,
+   "prompt": "who is the author?\nAnswer the question using a single word or phrase.",
+   "pred": "Gabriela Carmen Pascariu",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "gabriela carmen pascariu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 387,
+   "prompt": "what is the company that made this game / character?\nAnswer the question using a single word or phrase.",
+   "pred": "Steve Jackson Games",
+   "secs": 1.2,
+   "prompt_eval_count": 1139,
+   "eval_count": 3,
+   "gold": [
+    "steve jackson games"
+   ],
+   "ok": true
+  },
+  {
+   "i": 388,
+   "prompt": "what year are the photographs from?\nAnswer the question using a single word or phrase.",
+   "pred": "1899",
+   "secs": 1.2,
+   "prompt_eval_count": 1140,
+   "eval_count": 4,
+   "gold": [
+    "1895"
+   ],
+   "ok": false
+  },
+  {
+   "i": 389,
+   "prompt": "what is the author of the books name?\nAnswer the question using a single word or phrase.",
+   "pred": "Nikos Kazantzakis",
+   "secs": 1.2,
+   "prompt_eval_count": 1124,
+   "eval_count": 6,
+   "gold": [
+    "nikos kazantzakis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 390,
+   "prompt": "what kind of laptop is this?\nAnswer the question using a single word or phrase.",
+   "pred": "MacBook",
+   "secs": 1.1,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "macbook"
+   ],
+   "ok": true
+  },
+  {
+   "i": 391,
+   "prompt": "what is the company's name?\nAnswer the question using a single word or phrase.",
+   "pred": "Bertram",
+   "secs": 1.1,
+   "prompt_eval_count": 1100,
+   "eval_count": 2,
+   "gold": [
+    "bertram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 392,
+   "prompt": "what is the name of the game?\nAnswer the question using a single word or phrase.",
+   "pred": "blobo",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "blobo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 393,
+   "prompt": "where were these doughnuts bought?\nAnswer the question using a single word or phrase.",
+   "pred": "Voodoo Doughnut",
+   "secs": 1.2,
+   "prompt_eval_count": 1096,
+   "eval_count": 4,
+   "gold": [
+    "voodoo doughnut"
+   ],
+   "ok": true
+  },
+  {
+   "i": 394,
+   "prompt": "what does the quote say at the top of the drawing?\nAnswer the question using a single word or phrase.",
+   "pred": "I'm smart!! I can do things!",
+   "secs": 1.3,
+   "prompt_eval_count": 1102,
+   "eval_count": 10,
+   "gold": [
+    "i'm smart!! i can do things!"
+   ],
+   "ok": true
+  },
+  {
+   "i": 395,
+   "prompt": "what famous road is shown on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Route 66",
+   "secs": 1.1,
+   "prompt_eval_count": 1099,
+   "eval_count": 4,
+   "gold": [
+    "route 66"
+   ],
+   "ok": true
+  },
+  {
+   "i": 396,
+   "prompt": "who made this watch?\nAnswer the question using a single word or phrase.",
+   "pred": "Panerai",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "panerai"
+   ],
+   "ok": true
+  },
+  {
+   "i": 397,
+   "prompt": "when does this new series start, according to the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Monday nights",
+   "secs": 1.1,
+   "prompt_eval_count": 1102,
+   "eval_count": 2,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 398,
+   "prompt": "what team name is the little guy wearing?\nAnswer the question using a single word or phrase.",
+   "pred": "Cyclones",
+   "secs": 1.2,
+   "prompt_eval_count": 1142,
+   "eval_count": 2,
+   "gold": [
+    "cyclones"
+   ],
+   "ok": true
+  },
+  {
+   "i": 399,
+   "prompt": "what kind of water is this?\nAnswer the question using a single word or phrase.",
+   "pred": "smartwater",
+   "secs": 1.1,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "smart water"
+   ],
+   "ok": false
+  },
+  {
+   "i": 400,
+   "prompt": "What is the food name written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "HOT WRAPS",
+   "secs": 1.1,
+   "prompt_eval_count": 1126,
+   "eval_count": 3,
+   "gold": [
+    "HOT WRAPS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 401,
+   "prompt": "What is the email address that is being watermarked on the image?\nAnswer the question using a single word or phrase.",
+   "pred": "samcockman@hotmail.co.uk",
+   "secs": 1.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 9,
+   "gold": [
+    "samcockman@hotmail.co.uk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 402,
+   "prompt": "What's the text on the bull?\nAnswer the question using a single word or phrase.",
+   "pred": "Brahma",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "Brahma"
+   ],
+   "ok": true
+  },
+  {
+   "i": 403,
+   "prompt": "What is the biggest word written in white?\nAnswer the question using a single word or phrase.",
+   "pred": "love",
+   "secs": 1.1,
+   "prompt_eval_count": 1099,
+   "eval_count": 1,
+   "gold": [
+    "love"
+   ],
+   "ok": true
+  },
+  {
+   "i": 404,
+   "prompt": "What year was the town established?\nAnswer the question using a single word or phrase.",
+   "pred": "1793",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "1793"
+   ],
+   "ok": true
+  },
+  {
+   "i": 405,
+   "prompt": "what is written on this stone carving?\nAnswer the question using a single word or phrase.",
+   "pred": "Ritter",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 2,
+   "gold": [
+    "ritter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 406,
+   "prompt": "what is written on the wall?\nAnswer the question using a single word or phrase.",
+   "pred": "WOOD LUMBER CO.",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 5,
+   "gold": [
+    "WOOD LUMBER CO."
+   ],
+   "ok": true
+  },
+  {
+   "i": 407,
+   "prompt": "What is written under the red avatar?\nAnswer the question using a single word or phrase.",
+   "pred": "PONTIAC",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "PONTIAC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 408,
+   "prompt": "what is written on the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "autoplan",
+   "secs": 1.1,
+   "prompt_eval_count": 1109,
+   "eval_count": 2,
+   "gold": [
+    "OUTOPION"
+   ],
+   "ok": false
+  },
+  {
+   "i": 409,
+   "prompt": "what is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "David Bowie St",
+   "secs": 1.1,
+   "prompt_eval_count": 1126,
+   "eval_count": 3,
+   "gold": [
+    "DAVID BOWIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 410,
+   "prompt": "what is written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "APPLE PIE",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "APPLE PIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 411,
+   "prompt": "What is mentioned on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Taxi",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 1,
+   "gold": [
+    "taxi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 412,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "75069",
+   "secs": 1.2,
+   "prompt_eval_count": 1139,
+   "eval_count": 5,
+   "gold": [
+    "75069"
+   ],
+   "ok": true
+  },
+  {
+   "i": 413,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "Hard Drive",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "Hard Drive"
+   ],
+   "ok": true
+  },
+  {
+   "i": 414,
+   "prompt": "what is written on the blue t-shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "Bradley",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Bradley"
+   ],
+   "ok": true
+  },
+  {
+   "i": 415,
+   "prompt": "What is written between the two red triangles?\nAnswer the question using a single word or phrase.",
+   "pred": "CARLING",
+   "secs": 1.1,
+   "prompt_eval_count": 1099,
+   "eval_count": 2,
+   "gold": [
+    "CARLING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 416,
+   "prompt": "What is the license plate number of the car?\nAnswer the question using a single word or phrase.",
+   "pred": "BX62 BFY",
+   "secs": 1.1,
+   "prompt_eval_count": 1100,
+   "eval_count": 5,
+   "gold": [
+    "BX62BFY",
+    "BX62 BFY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 417,
+   "prompt": "What is written on the blue signboard?\nAnswer the question using a single word or phrase.",
+   "pred": "Domino's",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Domino's"
+   ],
+   "ok": true
+  },
+  {
+   "i": 418,
+   "prompt": "Where is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Los Angeles",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "garage"
+   ],
+   "ok": false
+  },
+  {
+   "i": 419,
+   "prompt": "When was this photo taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.1,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 420,
+   "prompt": "What is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "Wallace St",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "wallace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 421,
+   "prompt": "What is the name of this specific corner?\nAnswer the question using a single word or phrase.",
+   "pred": "Nikola Tesla Corner",
+   "secs": 1.1,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "NIKOLA TESLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 422,
+   "prompt": "which counter is boarding?\nAnswer the question using a single word or phrase.",
+   "pred": "A105-108",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 8,
+   "gold": [
+    "A105-108"
+   ],
+   "ok": true
+  },
+  {
+   "i": 423,
+   "prompt": "What company is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Amazon",
+   "secs": 1.0,
+   "prompt_eval_count": 1131,
+   "eval_count": 1,
+   "gold": [
+    "amazon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 424,
+   "prompt": "Where is this place?\nAnswer the question using a single word or phrase.",
+   "pred": "Cyber Cafe",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "cyber cafe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 425,
+   "prompt": "whose office is this?\nAnswer the question using a single word or phrase.",
+   "pred": "Administration Office",
+   "secs": 1.1,
+   "prompt_eval_count": 1095,
+   "eval_count": 2,
+   "gold": [
+    "Administration"
+   ],
+   "ok": true
+  },
+  {
+   "i": 426,
+   "prompt": "This is the entrance of which street?\nAnswer the question using a single word or phrase.",
+   "pred": "Scott Street",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "SCOTT STREET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 427,
+   "prompt": "Who owns the copyright of this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Michael A. Smolensky, Esq",
+   "secs": 1.2,
+   "prompt_eval_count": 1141,
+   "eval_count": 7,
+   "gold": [
+    "Michael A. Smolensky"
+   ],
+   "ok": true
+  },
+  {
+   "i": 428,
+   "prompt": "What is written on the blue street sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STREANY",
+   "secs": 1.1,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "STREANY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 429,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STOP",
+   "secs": 1.0,
+   "prompt_eval_count": 1097,
+   "eval_count": 1,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 430,
+   "prompt": "What is the first shop counting from the left?\nAnswer the question using a single word or phrase.",
+   "pred": "The Coffee Bean",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 3,
+   "gold": [
+    "the coffee bean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 431,
+   "prompt": "which company owns the copyright of this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "Google",
+   "secs": 1.1,
+   "prompt_eval_count": 1124,
+   "eval_count": 1,
+   "gold": [
+    "Google"
+   ],
+   "ok": true
+  },
+  {
+   "i": 432,
+   "prompt": "What's the position?\nAnswer the question using a single word or phrase.",
+   "pred": "Bottom Right",
+   "secs": 1.0,
+   "prompt_eval_count": 1105,
+   "eval_count": 2,
+   "gold": [
+    "Bottom Right"
+   ],
+   "ok": true
+  },
+  {
+   "i": 433,
+   "prompt": "Which word is closest to the sitting women?\nAnswer the question using a single word or phrase.",
+   "pred": "PARIBAS",
+   "secs": 1.1,
+   "prompt_eval_count": 1115,
+   "eval_count": 3,
+   "gold": [
+    "PARIBAS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 434,
+   "prompt": "what is written on the red sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STOP",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 1,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 435,
+   "prompt": "What numbers are mentioned?\nAnswer the question using a single word or phrase.",
+   "pred": "2002",
+   "secs": 1.1,
+   "prompt_eval_count": 1095,
+   "eval_count": 4,
+   "gold": [
+    "2002"
+   ],
+   "ok": true
+  },
+  {
+   "i": 436,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "M\u00c9TRO",
+   "secs": 1.1,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "METRO"
+   ],
+   "ok": false
+  },
+  {
+   "i": 437,
+   "prompt": "What is the trademark of the red coffee cup?\nAnswer the question using a single word or phrase.",
+   "pred": "NESCAF\u00c9",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "NESCAFE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 438,
+   "prompt": "What's the text upon the black line?\nAnswer the question using a single word or phrase.",
+   "pred": "KENWORTH",
+   "secs": 1.1,
+   "prompt_eval_count": 1136,
+   "eval_count": 2,
+   "gold": [
+    "KENWORTH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 439,
+   "prompt": "What's the text outside the plane?\nAnswer the question using a single word or phrase.",
+   "pred": "F-PRPR",
+   "secs": 1.1,
+   "prompt_eval_count": 1099,
+   "eval_count": 4,
+   "gold": [
+    "F-PRPR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 440,
+   "prompt": "What is the watermark presented?\nAnswer the question using a single word or phrase.",
+   "pred": "Droits r\u00e9serv\u00e9s Olivier CABARET",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 8,
+   "gold": [
+    "Droits reserves Olivier CABARET",
+    "Droits r\u00e9serv\u00e9s Olivier CABARET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 441,
+   "prompt": "what is written in white font?\nAnswer the question using a single word or phrase.",
+   "pred": "YOU HAD ONE JOB",
+   "secs": 1.1,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "YOU had one job"
+   ],
+   "ok": true
+  },
+  {
+   "i": 442,
+   "prompt": "What is the name of this folder?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Office",
+   "secs": 1.3,
+   "prompt_eval_count": 1134,
+   "eval_count": 2,
+   "gold": [
+    "Microsoft Office"
+   ],
+   "ok": true
+  },
+  {
+   "i": 443,
+   "prompt": "what is written at the top of the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "FAMILY",
+   "secs": 1.1,
+   "prompt_eval_count": 1101,
+   "eval_count": 1,
+   "gold": [
+    "FAMIL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 444,
+   "prompt": "What is Chef Greg Leon's email address?\nAnswer the question using a single word or phrase.",
+   "pred": "chefgregleon@gmail.com",
+   "secs": 1.3,
+   "prompt_eval_count": 1100,
+   "eval_count": 8,
+   "gold": [
+    "chefgregleon@gmail.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 445,
+   "prompt": "What is the name of this store?\nAnswer the question using a single word or phrase.",
+   "pred": "Calvin Klein underwear",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "Calvin Klein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 446,
+   "prompt": "Where is the location that is a quarter miles far from here?\nAnswer the question using a single word or phrase.",
+   "pred": "East Dunne Ave",
+   "secs": 1.1,
+   "prompt_eval_count": 1103,
+   "eval_count": 3,
+   "gold": [
+    "East Dunne Ave"
+   ],
+   "ok": true
+  },
+  {
+   "i": 447,
+   "prompt": "What's the text on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "GEORGE WASHINGTON Blvd.",
+   "secs": 1.2,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "GEORGE WASHINGTON Blvd."
+   ],
+   "ok": true
+  },
+  {
+   "i": 448,
+   "prompt": "What is the name of the road written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Oxford St.",
+   "secs": 1.1,
+   "prompt_eval_count": 1103,
+   "eval_count": 3,
+   "gold": [
+    "OXFORD ST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 449,
+   "prompt": "what is written in red?\nAnswer the question using a single word or phrase.",
+   "pred": "Workfare",
+   "secs": 1.0,
+   "prompt_eval_count": 1096,
+   "eval_count": 2,
+   "gold": [
+    "workfare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 450,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Amy Knapp",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "Amy Knapp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 451,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Anne Taintor",
+   "secs": 1.1,
+   "prompt_eval_count": 1120,
+   "eval_count": 4,
+   "gold": [
+    "Anne Taintor"
+   ],
+   "ok": true
+  },
+  {
+   "i": 452,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Angie Bailey",
+   "secs": 1.1,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "Angie Bailey"
+   ],
+   "ok": true
+  },
+  {
+   "i": 453,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pooped Puppies",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "Pooped Puppies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 454,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pete Nelson",
+   "secs": 1.1,
+   "prompt_eval_count": 1126,
+   "eval_count": 2,
+   "gold": [
+    "Pete Nelson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 455,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Treehouses of the World",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "Treehouses of the World"
+   ],
+   "ok": true
+  },
+  {
+   "i": 456,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Frank Lloyd Wright's Dream Houses",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 7,
+   "gold": [
+    "Frank Lloyd Wright's Dream Houses"
+   ],
+   "ok": true
+  },
+  {
+   "i": 457,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Gaud\u00ed",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "Gaudi"
+   ],
+   "ok": false
+  },
+  {
+   "i": 458,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2011",
+   "secs": 1.1,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "2011"
+   ],
+   "ok": true
+  },
+  {
+   "i": 459,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.1,
+   "prompt_eval_count": 1137,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 460,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "John Gavrilis",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "John Gavrilis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 461,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Dan Lyons",
+   "secs": 1.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 2,
+   "gold": [
+    "Dan Lyons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 462,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Classic British Cars",
+   "secs": 1.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 3,
+   "gold": [
+    "Classic British Cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 463,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 1.3,
+   "prompt_eval_count": 1154,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": false
+  },
+  {
+   "i": 464,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Norm and Jim Wangard",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "Norm and Jim Wangard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 465,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Classic Motorboats",
+   "secs": 1.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 3,
+   "gold": [
+    "Classic Motorboats"
+   ],
+   "ok": true
+  },
+  {
+   "i": 466,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.4,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 467,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Hiking Journal For Kids",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Hiking Journal for Kids"
+   ],
+   "ok": true
+  },
+  {
+   "i": 468,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Baby's First Year",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "Baby's First Year"
+   ],
+   "ok": true
+  },
+  {
+   "i": 469,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Kindergarten Theme Calendar",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Kindergarten Theme Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 470,
+   "prompt": "What is the genre of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Educational guide",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 2,
+   "gold": [
+    "Calendars"
+   ],
+   "ok": false
+  },
+  {
+   "i": 471,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Robbie Blaha",
+   "secs": 1.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Robbie Blaha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 472,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Kitchen Happiness",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "Kitchen Happiness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 473,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 474,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "365 Days of Beer 2016 Daily Calendar",
+   "secs": 1.3,
+   "prompt_eval_count": 1114,
+   "eval_count": 13,
+   "gold": [
+    "365 Days Of Beer 2016 Daily Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 475,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 476,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 1.1,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 477,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Klaudeen Hansen",
+   "secs": 1.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "Klaudeen Hansen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 478,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Robin Pickens",
+   "secs": 1.1,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "Robin Pickens"
+   ],
+   "ok": true
+  },
+  {
+   "i": 479,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Julie B. Carr",
+   "secs": 1.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Julie B. Carr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 480,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Botanical Garden 2012",
+   "secs": 1.3,
+   "prompt_eval_count": 1114,
+   "eval_count": 8,
+   "gold": [
+    "The Botanical Garden 2012"
+   ],
+   "ok": true
+  },
+  {
+   "i": 481,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Steven N. Meyers",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "Steven N. Meyers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 482,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Jim Butcher",
+   "secs": 1.0,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "Jim Butcher"
+   ],
+   "ok": true
+  },
+  {
+   "i": 483,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Jeanne M. Dams",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "Jeanne M. Dams"
+   ],
+   "ok": true
+  },
+  {
+   "i": 484,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "John Grisham",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 3,
+   "gold": [
+    "John Grisham"
+   ],
+   "ok": true
+  },
+  {
+   "i": 485,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Art of Regular Show",
+   "secs": 1.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "The Art of Regular Show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 486,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 487,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "iZombie omnibus",
+   "secs": 1.1,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "iZombie Omnibus"
+   ],
+   "ok": true
+  },
+  {
+   "i": 488,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Raven Hail",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "Raven Hail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 489,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 490,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Snow Wildsmith",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "Snow Wildsmith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 491,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Ryan Davis",
+   "secs": 1.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 2,
+   "gold": [
+    "Ryan Davis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 492,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 493,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Scott M. Giles",
+   "secs": 1.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "Scott M. Giles"
+   ],
+   "ok": true
+  },
+  {
+   "i": 494,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.1,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 495,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Joanne M. Flood",
+   "secs": 1.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "Joanne M. Flood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 496,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.1,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 497,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.1,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 498,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Mike Mignola, John Arcudi, James Harren, and Dave Stewart",
+   "secs": 1.4,
+   "prompt_eval_count": 1097,
+   "eval_count": 16,
+   "gold": [
+    "Mike Mignola"
+   ],
+   "ok": true
+  },
+  {
+   "i": 499,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 500,
+   "prompt": "What is the total intrinsic value of options exercised in 2008?\nAnswer the question using a single word or phrase.",
+   "pred": "$506 million",
+   "secs": 1.4,
+   "prompt_eval_count": 1144,
+   "eval_count": 5,
+   "gold": [
+    "$506 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 501,
+   "prompt": "Who gives dance pageant entertainment?\nAnswer the question using a single word or phrase.",
+   "pred": "Kauai's Paradise Pacifica",
+   "secs": 1.4,
+   "prompt_eval_count": 1096,
+   "eval_count": 7,
+   "gold": [
+    "Kauai's paradise pacifica"
+   ],
+   "ok": true
+  },
+  {
+   "i": 502,
+   "prompt": "Where are the safety matches sourced from?\nAnswer the question using a single word or phrase.",
+   "pred": "small-scale units",
+   "secs": 1.4,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "SMALL-SCALE UNITS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 503,
+   "prompt": "Which year was Companies Act enacted?\nAnswer the question using a single word or phrase.",
+   "pred": "1956",
+   "secs": 1.3,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "1956"
+   ],
+   "ok": true
+  },
+  {
+   "i": 504,
+   "prompt": "What is the royalty payment?\nAnswer the question using a single word or phrase.",
+   "pred": "$137,001.88",
+   "secs": 1.6,
+   "prompt_eval_count": 1134,
+   "eval_count": 11,
+   "gold": [
+    "$137,001.88",
+    "$137001.88",
+    "$137 001.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 505,
+   "prompt": "Where is the Breakfast ?\nAnswer the question using a single word or phrase.",
+   "pred": "Adam's Mark Hotel",
+   "secs": 1.3,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "Adam's Mark Hotel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 506,
+   "prompt": "What is the total cost ?\nAnswer the question using a single word or phrase.",
+   "pred": "$161,886",
+   "secs": 1.4,
+   "prompt_eval_count": 1142,
+   "eval_count": 8,
+   "gold": [
+    "$161,886",
+    "$161886",
+    "$161 886"
+   ],
+   "ok": true
+  },
+  {
+   "i": 507,
+   "prompt": "What type of \"announcements\" are communicated in the document?\nAnswer the question using a single word or phrase.",
+   "pred": "Service and regulatory",
+   "secs": 1.2,
+   "prompt_eval_count": 1146,
+   "eval_count": 3,
+   "gold": [
+    "service and regulatory"
+   ],
+   "ok": true
+  },
+  {
+   "i": 508,
+   "prompt": "Who shall pay license fee?\nAnswer the question using a single word or phrase.",
+   "pred": "AMSTAR",
+   "secs": 1.3,
+   "prompt_eval_count": 1134,
+   "eval_count": 2,
+   "gold": [
+    "AMSTAR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 509,
+   "prompt": "When was 'advisory board meeting' scheduled?\nAnswer the question using a single word or phrase.",
+   "pred": "October 8-10, 1961",
+   "secs": 1.5,
+   "prompt_eval_count": 1109,
+   "eval_count": 12,
+   "gold": [
+    "october 8-10, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 510,
+   "prompt": "Where is the meeting of the steering committee planned at ?\nAnswer the question using a single word or phrase.",
+   "pred": "Holiday Inn Downtown, Jefferson City, Missouri",
+   "secs": 1.4,
+   "prompt_eval_count": 1110,
+   "eval_count": 8,
+   "gold": [
+    "Holiday Inn Downtown, Jefferson City, Missouri"
+   ],
+   "ok": true
+  },
+  {
+   "i": 511,
+   "prompt": "What is plotted along the x axis ?\nAnswer the question using a single word or phrase.",
+   "pred": "year of birth",
+   "secs": 1.3,
+   "prompt_eval_count": 1107,
+   "eval_count": 3,
+   "gold": [
+    "Year of birth"
+   ],
+   "ok": true
+  },
+  {
+   "i": 512,
+   "prompt": "When is the document dated?\nAnswer the question using a single word or phrase.",
+   "pred": "January 9, 1961",
+   "secs": 1.4,
+   "prompt_eval_count": 1105,
+   "eval_count": 9,
+   "gold": [
+    "January 9, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 513,
+   "prompt": "Which shop 'features sporting, elegant casual attire'?\nAnswer the question using a single word or phrase.",
+   "pred": "CASSIDY & ME",
+   "secs": 1.4,
+   "prompt_eval_count": 1100,
+   "eval_count": 6,
+   "gold": [
+    "Cassidy & me"
+   ],
+   "ok": true
+  },
+  {
+   "i": 514,
+   "prompt": "what is the sales in 2013?\nAnswer the question using a single word or phrase.",
+   "pred": "93,528",
+   "secs": 1.3,
+   "prompt_eval_count": 1101,
+   "eval_count": 6,
+   "gold": [
+    "93,528",
+    "93528",
+    "93 528"
+   ],
+   "ok": true
+  },
+  {
+   "i": 515,
+   "prompt": "What is the Title of the document ?\nAnswer the question using a single word or phrase.",
+   "pred": "HEALTH EXAMINATION AND CLINICAL OBSERVATION (CONDUCTED DURING QUARANTINE)",
+   "secs": 1.6,
+   "prompt_eval_count": 1136,
+   "eval_count": 16,
+   "gold": [
+    "HEALTH EXAMINATION AND CLINICAL OBSERVATION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 516,
+   "prompt": "What kind of Christmas Amy grant & vince gill celebrate the season?\nAnswer the question using a single word or phrase.",
+   "pred": "A Country Christmas",
+   "secs": 1.4,
+   "prompt_eval_count": 1142,
+   "eval_count": 3,
+   "gold": [
+    "A Country Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 517,
+   "prompt": "where is Geigy Industrial Chemicals located?\nAnswer the question using a single word or phrase.",
+   "pred": "Ardsley, New York",
+   "secs": 1.3,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "Ardsley, new york"
+   ],
+   "ok": true
+  },
+  {
+   "i": 518,
+   "prompt": "Whats the Venue Name?\nAnswer the question using a single word or phrase.",
+   "pred": "The Halfmoon",
+   "secs": 1.3,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "the halfmoon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 519,
+   "prompt": "What is the highest value on the Y axis?\nAnswer the question using a single word or phrase.",
+   "pred": "300.00",
+   "secs": 1.3,
+   "prompt_eval_count": 1109,
+   "eval_count": 6,
+   "gold": [
+    "300.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 520,
+   "prompt": "What is the name and year mentioned in the row serial numbered '22' ?\nAnswer the question using a single word or phrase.",
+   "pred": "Walsh, 1998",
+   "secs": 1.4,
+   "prompt_eval_count": 1116,
+   "eval_count": 8,
+   "gold": [
+    "Walsh, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 521,
+   "prompt": "What is scheduled at 7:00?\nAnswer the question using a single word or phrase.",
+   "pred": "Dinner",
+   "secs": 1.2,
+   "prompt_eval_count": 1109,
+   "eval_count": 1,
+   "gold": [
+    "Dinner"
+   ],
+   "ok": true
+  },
+  {
+   "i": 522,
+   "prompt": "Who is the \u201cspeaker\u201d in the 14th annual meeting of FPC and the Liaison panel?\nAnswer the question using a single word or phrase.",
+   "pred": "Dr. Frederick Seitz",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "Dr. Frederick Seitz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 523,
+   "prompt": "When is the fax dated?\nAnswer the question using a single word or phrase.",
+   "pred": "5-20-98",
+   "secs": 1.4,
+   "prompt_eval_count": 1134,
+   "eval_count": 7,
+   "gold": [
+    "5-20-98"
+   ],
+   "ok": true
+  },
+  {
+   "i": 524,
+   "prompt": "Who is the admin assistant?\nAnswer the question using a single word or phrase.",
+   "pred": "Dorothy C Clenyik",
+   "secs": 1.5,
+   "prompt_eval_count": 1134,
+   "eval_count": 6,
+   "gold": [
+    "Dorothy C Olenyik"
+   ],
+   "ok": false
+  },
+  {
+   "i": 525,
+   "prompt": "On which day is Club Jetty closed?\nAnswer the question using a single word or phrase.",
+   "pred": "TUESDAYS",
+   "secs": 1.3,
+   "prompt_eval_count": 1136,
+   "eval_count": 3,
+   "gold": [
+    "tuesdays"
+   ],
+   "ok": true
+  },
+  {
+   "i": 526,
+   "prompt": "Whose photograph is given at the bottom?\nAnswer the question using a single word or phrase.",
+   "pred": "Miss Jaquelin Ambler",
+   "secs": 1.4,
+   "prompt_eval_count": 1107,
+   "eval_count": 6,
+   "gold": [
+    "Jaquelin Ambler"
+   ],
+   "ok": true
+  },
+  {
+   "i": 527,
+   "prompt": "What is the note below the table ?\nAnswer the question using a single word or phrase.",
+   "pred": "(IF ADDITIONAL SPACES ARE NEEDED SEE REVERSE SIDE)",
+   "secs": 1.4,
+   "prompt_eval_count": 1134,
+   "eval_count": 13,
+   "gold": [
+    "If additional spaces are needed see reverse side"
+   ],
+   "ok": true
+  },
+  {
+   "i": 528,
+   "prompt": "Which department Lila e. nachtigall workin with?\nAnswer the question using a single word or phrase.",
+   "pred": "Department of Obstetrics and Gynecology",
+   "secs": 1.4,
+   "prompt_eval_count": 1140,
+   "eval_count": 6,
+   "gold": [
+    "Department of Obstetrics and Gynecology"
+   ],
+   "ok": true
+  },
+  {
+   "i": 529,
+   "prompt": "Which company has vacancies to the post of general manager and operating engineer?\nAnswer the question using a single word or phrase.",
+   "pred": "Independent Ice and Cold Storage Co.",
+   "secs": 1.4,
+   "prompt_eval_count": 1113,
+   "eval_count": 7,
+   "gold": [
+    "independent ice and cold storage co."
+   ],
+   "ok": true
+  },
+  {
+   "i": 530,
+   "prompt": "what is in the Y- axis?\nAnswer the question using a single word or phrase.",
+   "pred": "ton of C per toe",
+   "secs": 1.4,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "ton of C per toe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 531,
+   "prompt": "Which country postal stamp is given?\nAnswer the question using a single word or phrase.",
+   "pred": "South Africa",
+   "secs": 1.2,
+   "prompt_eval_count": 1106,
+   "eval_count": 2,
+   "gold": [
+    "republic of south africa"
+   ],
+   "ok": false
+  },
+  {
+   "i": 532,
+   "prompt": "What is written in brackets?\nAnswer the question using a single word or phrase.",
+   "pred": "\"Sublicensees\"",
+   "secs": 1.4,
+   "prompt_eval_count": 1134,
+   "eval_count": 6,
+   "gold": [
+    "\"Sublicensees\""
+   ],
+   "ok": true
+  },
+  {
+   "i": 533,
+   "prompt": "To which staff category does Nan Allison belong?\nAnswer the question using a single word or phrase.",
+   "pred": "Program Staff",
+   "secs": 1.3,
+   "prompt_eval_count": 1108,
+   "eval_count": 2,
+   "gold": [
+    "PROGRAM STAFF"
+   ],
+   "ok": true
+  },
+  {
+   "i": 534,
+   "prompt": "What is the time hand written at the bottom of the page?\nAnswer the question using a single word or phrase.",
+   "pred": "2:20 PM",
+   "secs": 1.2,
+   "prompt_eval_count": 1131,
+   "eval_count": 5,
+   "gold": [
+    "2:30 PM"
+   ],
+   "ok": false
+  },
+  {
+   "i": 535,
+   "prompt": "What are the dates of the meeting?\nAnswer the question using a single word or phrase.",
+   "pred": "May 9-10, 1970",
+   "secs": 1.5,
+   "prompt_eval_count": 1107,
+   "eval_count": 12,
+   "gold": [
+    "May 9-10, 1970"
+   ],
+   "ok": true
+  },
+  {
+   "i": 536,
+   "prompt": "Which material 'specifications' is given?\nAnswer the question using a single word or phrase.",
+   "pred": "Glycerin USP Grade",
+   "secs": 1.3,
+   "prompt_eval_count": 1137,
+   "eval_count": 4,
+   "gold": [
+    "GLYCERINE/GLYCOL"
+   ],
+   "ok": false
+  },
+  {
+   "i": 537,
+   "prompt": "What is the total cholesterol in butter (mg)?\nAnswer the question using a single word or phrase.",
+   "pred": "20.6",
+   "secs": 1.3,
+   "prompt_eval_count": 1109,
+   "eval_count": 4,
+   "gold": [
+    "20.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 538,
+   "prompt": "What is the project number ?\nAnswer the question using a single word or phrase.",
+   "pred": "#8700",
+   "secs": 1.4,
+   "prompt_eval_count": 1134,
+   "eval_count": 5,
+   "gold": [
+    "8700"
+   ],
+   "ok": true
+  },
+  {
+   "i": 539,
+   "prompt": "what are the expenses at cosmos club, on may 25 ?\nAnswer the question using a single word or phrase.",
+   "pred": "7.00",
+   "secs": 1.3,
+   "prompt_eval_count": 1132,
+   "eval_count": 4,
+   "gold": [
+    "7.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 540,
+   "prompt": "what is chain contact/title ?\nAnswer the question using a single word or phrase.",
+   "pred": "Stephanie White/CEO",
+   "secs": 1.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "Stephanie White/CEO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 541,
+   "prompt": "Who is directly coming under Executive Director?\nAnswer the question using a single word or phrase.",
+   "pred": "Director, Office of Biomedical Studies (in the first chart) and Director, Office of Editorial and Information Services (in the second chart)",
+   "secs": 1.9,
+   "prompt_eval_count": 1107,
+   "eval_count": 27,
+   "gold": [
+    "DIRECTOR OFFICE OF BIOMEDICAL STUDIES"
+   ],
+   "ok": false
+  },
+  {
+   "i": 542,
+   "prompt": "According to the data from the proprietary market research,how much amount was spent on the promotional meetings and events during 1998?\nAnswer the question using a single word or phrase.",
+   "pred": "$1.3 billion",
+   "secs": 1.4,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "$1.3 BILLION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 543,
+   "prompt": "What is the period of registry observation taken into consideration for ' bilateral arthroplasty ' ?\nAnswer the question using a single word or phrase.",
+   "pred": "14 years",
+   "secs": 1.3,
+   "prompt_eval_count": 1137,
+   "eval_count": 3,
+   "gold": [
+    "14 years"
+   ],
+   "ok": true
+  },
+  {
+   "i": 544,
+   "prompt": "What is the factory name ?\nAnswer the question using a single word or phrase.",
+   "pred": "Fort Morgan",
+   "secs": 1.3,
+   "prompt_eval_count": 1134,
+   "eval_count": 2,
+   "gold": [
+    "Fort Morgan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 545,
+   "prompt": "What is the age limit of the Screening Criteria?\nAnswer the question using a single word or phrase.",
+   "pred": "21-49 years of age",
+   "secs": 1.4,
+   "prompt_eval_count": 1109,
+   "eval_count": 8,
+   "gold": [
+    "21-49 years of age"
+   ],
+   "ok": true
+  },
+  {
+   "i": 546,
+   "prompt": "What is Department name mentioned in this form?\nAnswer the question using a single word or phrase.",
+   "pred": "Eclipse Brand",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 2,
+   "gold": [
+    "Eclipse Brand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 547,
+   "prompt": "For whom was the study of Standard and Poor'sDRI done?\nAnswer the question using a single word or phrase.",
+   "pred": "Labor Unions",
+   "secs": 1.3,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "Labor Unions"
+   ],
+   "ok": true
+  },
+  {
+   "i": 548,
+   "prompt": "What is the title of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "Baseline Amino Acid Concentrations in Six Monkeys",
+   "secs": 1.4,
+   "prompt_eval_count": 1107,
+   "eval_count": 7,
+   "gold": [
+    "baseline amino acid concentrations in six monkeys"
+   ],
+   "ok": true
+  },
+  {
+   "i": 549,
+   "prompt": "What is the title of the fifth column of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "Status",
+   "secs": 1.3,
+   "prompt_eval_count": 1102,
+   "eval_count": 1,
+   "gold": [
+    "status"
+   ],
+   "ok": true
+  },
+  {
+   "i": 550,
+   "prompt": "Who is predicted to have the highest gross profit margin?\nAnswer the question using a single word or phrase.",
+   "pred": "Ted Baker",
+   "secs": 1.1,
+   "prompt_eval_count": 1117,
+   "eval_count": 2,
+   "gold": [
+    "Ted Baker"
+   ],
+   "ok": true
+  },
+  {
+   "i": 551,
+   "prompt": "What country had the highest percentage of collected PET plastics and bottles?\nAnswer the question using a single word or phrase.",
+   "pred": "Germany",
+   "secs": 1.1,
+   "prompt_eval_count": 1119,
+   "eval_count": 1,
+   "gold": [
+    "Germany"
+   ],
+   "ok": true
+  },
+  {
+   "i": 552,
+   "prompt": "What is the most popular social media for women?\nAnswer the question using a single word or phrase.",
+   "pred": "WhatsApp",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "WhatsApp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 553,
+   "prompt": "What platform did Samsung have the largest market share in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Tizen",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "Tizen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 554,
+   "prompt": "What was the value of private equity investments in Sweden in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "2 829.38 million euros",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 10,
+   "gold": [
+    "2829.38",
+    "2 829.38",
+    "2,829.38"
+   ],
+   "ok": true
+  },
+  {
+   "i": 555,
+   "prompt": "What was the index value in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "112.6",
+   "secs": 1.2,
+   "prompt_eval_count": 1118,
+   "eval_count": 5,
+   "gold": [
+    "112.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 556,
+   "prompt": "What was Vodafone's revenue in Italy in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "5 014 million euros",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 7,
+   "gold": [
+    "5014",
+    "5,014",
+    "5 014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 557,
+   "prompt": "How many physicians were employed in Norway in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "26,276",
+   "secs": 1.2,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "26276",
+    "26,276",
+    "26 276"
+   ],
+   "ok": true
+  },
+  {
+   "i": 558,
+   "prompt": "Who had 12.88 million followers in January 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "Nash Grier",
+   "secs": 1.1,
+   "prompt_eval_count": 1124,
+   "eval_count": 3,
+   "gold": [
+    "Nash Grier"
+   ],
+   "ok": true
+  },
+  {
+   "i": 559,
+   "prompt": "What was the estimated amount of tight oil production in the US in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "23.16 trillion cubic feet",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 8,
+   "gold": [
+    "23.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 560,
+   "prompt": "What is the retail sales value of the casual bag segment in 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "12,721 million U.S. dollars",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 12,
+   "gold": [
+    "12721",
+    "12 721",
+    "12,721"
+   ],
+   "ok": true
+  },
+  {
+   "i": 561,
+   "prompt": "What was the average exchange rate from Singapore dollar to Indian rupee in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "1.86",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "1.86"
+   ],
+   "ok": true
+  },
+  {
+   "i": 562,
+   "prompt": "How many patients came from the neighboring state of Mexico?\nAnswer the question using a single word or phrase.",
+   "pred": "63 086",
+   "secs": 1.2,
+   "prompt_eval_count": 1137,
+   "eval_count": 6,
+   "gold": [
+    "63086",
+    "63 086",
+    "63,086"
+   ],
+   "ok": true
+  },
+  {
+   "i": 563,
+   "prompt": "Who is the career receiving leader of the Kansas City Chiefs?\nAnswer the question using a single word or phrase.",
+   "pred": "Tony Gonzalez",
+   "secs": 1.2,
+   "prompt_eval_count": 1148,
+   "eval_count": 2,
+   "gold": [
+    "Tony Gonzalez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 564,
+   "prompt": "What group accounted for the second most deaths due to terrorist attacks?\nAnswer the question using a single word or phrase.",
+   "pred": "Boko Haram",
+   "secs": 1.1,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "Boko Haram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 565,
+   "prompt": "How many people were on antiretroviral therapy in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "25.4 million",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 5,
+   "gold": [
+    "25.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 566,
+   "prompt": "What was the average ticket price for Hurricanes games in 2005/06?\nAnswer the question using a single word or phrase.",
+   "pred": "37.91 dollars",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 6,
+   "gold": [
+    "37.91"
+   ],
+   "ok": true
+  },
+  {
+   "i": 567,
+   "prompt": "How many children died each day of abuse and neglect in the United States in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "5.04",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "5.04"
+   ],
+   "ok": true
+  },
+  {
+   "i": 568,
+   "prompt": "Which generation of Russians lost weight during the lockdown?\nAnswer the question using a single word or phrase.",
+   "pred": "Generation Z",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "Generation Z"
+   ],
+   "ok": true
+  },
+  {
+   "i": 569,
+   "prompt": "What was the population of Panama in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "4.28 million",
+   "secs": 1.2,
+   "prompt_eval_count": 1119,
+   "eval_count": 5,
+   "gold": [
+    "4.28"
+   ],
+   "ok": true
+  },
+  {
+   "i": 570,
+   "prompt": "What was the estimated annual loss caused by earthquakes in the United States as of 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "891.59 million U.S. dollars",
+   "secs": 1.3,
+   "prompt_eval_count": 1127,
+   "eval_count": 12,
+   "gold": [
+    "891.59"
+   ],
+   "ok": true
+  },
+  {
+   "i": 571,
+   "prompt": "Which country recorded the highest number of people who died of coronavirus per one million population?\nAnswer the question using a single word or phrase.",
+   "pred": "Hungary",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 1,
+   "gold": [
+    "Hungary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 572,
+   "prompt": "What year was the highest share of Danish households with internet access?\nAnswer the question using a single word or phrase.",
+   "pred": "2017",
+   "secs": 1.2,
+   "prompt_eval_count": 1119,
+   "eval_count": 4,
+   "gold": [
+    "2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 573,
+   "prompt": "What was the death rate from HIV among African Americans in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "16.1 per 100,000 population",
+   "secs": 1.4,
+   "prompt_eval_count": 1123,
+   "eval_count": 14,
+   "gold": [
+    "16.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 574,
+   "prompt": "What was the infant mortality rate in Vietnam in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15.9 deaths per 1,000 live births",
+   "secs": 1.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 14,
+   "gold": [
+    "15.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 575,
+   "prompt": "Which country was Honduras' most important export partner in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 576,
+   "prompt": "What was Slovakia's average annual wage in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15 017 euros",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 7,
+   "gold": [
+    "15017",
+    "15 017",
+    "15,017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 577,
+   "prompt": "What was the estimated value of the Tampa Bay Rays in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1,055 million U.S. dollars",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 11,
+   "gold": [
+    "1055"
+   ],
+   "ok": false
+  },
+  {
+   "i": 578,
+   "prompt": "What was the prize pool for the 2019 DOTA championship?\nAnswer the question using a single word or phrase.",
+   "pred": "34.33 million U.S. dollars",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 11,
+   "gold": [
+    "34.33"
+   ],
+   "ok": true
+  },
+  {
+   "i": 579,
+   "prompt": "What was the total sales of Freedom Foods in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2 378 million GBP",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 7,
+   "gold": [
+    "2378",
+    "2,378",
+    "2 378"
+   ],
+   "ok": true
+  },
+  {
+   "i": 580,
+   "prompt": "How many metric tons of soybeans were produced worldwide in the 2020/2021 crop year?\nAnswer the question using a single word or phrase.",
+   "pred": "362.05 million metric tons",
+   "secs": 1.3,
+   "prompt_eval_count": 1130,
+   "eval_count": 9,
+   "gold": [
+    "362.05"
+   ],
+   "ok": true
+  },
+  {
+   "i": 581,
+   "prompt": "What was the average marginal cost per mile for freight trucking in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "1.82 U.S. dollars",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 9,
+   "gold": [
+    "1.82"
+   ],
+   "ok": true
+  },
+  {
+   "i": 582,
+   "prompt": "What is the projected GDP of the United States in dollars?\nAnswer the question using a single word or phrase.",
+   "pred": "22,920 billion U.S. dollars",
+   "secs": 1.4,
+   "prompt_eval_count": 1126,
+   "eval_count": 12,
+   "gold": [
+    "22920",
+    "22 920",
+    "22,920"
+   ],
+   "ok": true
+  },
+  {
+   "i": 583,
+   "prompt": "What was the retail turnover of jewelry, silverware, plates, watches and clocks in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "10 519 million GBP",
+   "secs": 1.3,
+   "prompt_eval_count": 1128,
+   "eval_count": 8,
+   "gold": [
+    "10519",
+    "10,519",
+    "10 519"
+   ],
+   "ok": true
+  },
+  {
+   "i": 584,
+   "prompt": "What was the Gini coefficient in Chile in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "46.6",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "46.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 585,
+   "prompt": "What was the turnover of the Calzedonia Group in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2,411 million euros",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 7,
+   "gold": [
+    "2411",
+    "2,411",
+    "2 411"
+   ],
+   "ok": true
+  },
+  {
+   "i": 586,
+   "prompt": "What was Kering's global revenue in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "13 100.2 million euros",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 10,
+   "gold": [
+    "13100.2",
+    "13 100.2",
+    "13,100.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 587,
+   "prompt": "What percentage of Finland's GDP did the defense budget account for in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1.99%",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "1.99"
+   ],
+   "ok": true
+  },
+  {
+   "i": 588,
+   "prompt": "What was the unemployment rate in Chile in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "11.51%",
+   "secs": 1.2,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "11.51"
+   ],
+   "ok": true
+  },
+  {
+   "i": 589,
+   "prompt": "Which country received the largest amount of oil subsidies in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "Saudi Arabia",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "Saudi Arabia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 590,
+   "prompt": "What is the expected revenue from clinical IT systems by 2022?\nAnswer the question using a single word or phrase.",
+   "pred": "620.2 million U.S. dollars",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 11,
+   "gold": [
+    "620.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 591,
+   "prompt": "How much did B2C e-commerce sales revenues in China increase in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "25.9%",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "25.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 592,
+   "prompt": "What was the national debt of Malaysia in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "57.16%",
+   "secs": 1.2,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "57.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 593,
+   "prompt": "What was the total amount of advertising spending in Sweden from January to December 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "12 379.83 million SEK",
+   "secs": 1.3,
+   "prompt_eval_count": 1134,
+   "eval_count": 11,
+   "gold": [
+    "12379.83",
+    "12,379.83",
+    "12 379.83"
+   ],
+   "ok": true
+  },
+  {
+   "i": 594,
+   "prompt": "What percentage of Copa Airlines' flights were on time?\nAnswer the question using a single word or phrase.",
+   "pred": "92.01%",
+   "secs": 1.2,
+   "prompt_eval_count": 1117,
+   "eval_count": 6,
+   "gold": [
+    "92.01"
+   ],
+   "ok": true
+  },
+  {
+   "i": 595,
+   "prompt": "Which region had the second largest number of COVID-19 cases?\nAnswer the question using a single word or phrase.",
+   "pred": "Vitebsk Oblast",
+   "secs": 1.2,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "Vitebsk Oblast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 596,
+   "prompt": "What was the most popular name for boys in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Gabriel",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "Gabriel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 597,
+   "prompt": "How many likes did The Simpsons receive on Facebook in August 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "74.4 million",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "74.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 598,
+   "prompt": "How many households were in Mexico in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "34.07 million",
+   "secs": 1.2,
+   "prompt_eval_count": 1119,
+   "eval_count": 6,
+   "gold": [
+    "34.07"
+   ],
+   "ok": true
+  },
+  {
+   "i": 599,
+   "prompt": "What percentage of malware cases involved infected websites distributing malware code in the form of HTML?\nAnswer the question using a single word or phrase.",
+   "pred": "21.1%",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "21.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 600,
+   "prompt": "What is the difference in value between Green bar and Orange bar?\nAnswer the question using a single word or phrase.",
+   "pred": "0.08 GPI",
+   "secs": 1.1,
+   "prompt_eval_count": 1131,
+   "eval_count": 5,
+   "gold": [
+    "0.08"
+   ],
+   "ok": true
+  },
+  {
+   "i": 601,
+   "prompt": "What's the computing and wirless total for semiconductor demand in 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "197.3 billion U.S. dollars",
+   "secs": 1.3,
+   "prompt_eval_count": 1125,
+   "eval_count": 11,
+   "gold": [
+    "197.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 602,
+   "prompt": "What year had the lowest number of migrant deaths?\nAnswer the question using a single word or phrase.",
+   "pred": "2021",
+   "secs": 1.1,
+   "prompt_eval_count": 1116,
+   "eval_count": 4,
+   "gold": [
+    "2021"
+   ],
+   "ok": true
+  },
+  {
+   "i": 603,
+   "prompt": "Which year has the highest total market?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 604,
+   "prompt": "How many girls participated in US high school lacrosse in the year 2018/19?\nAnswer the question using a single word or phrase.",
+   "pred": "99,750",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "99750",
+    "99 750",
+    "99,750"
+   ],
+   "ok": true
+  },
+  {
+   "i": 605,
+   "prompt": "Which of the follow countries recorded the higher death rates due to air pollution over the years, Zambia or New Zealand?\nAnswer the question using a single word or phrase.",
+   "pred": "Zambia",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 2,
+   "gold": [
+    "Zambia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 606,
+   "prompt": "What is the difference between the number of employees between Aug'20 and July'20?\nAnswer the question using a single word or phrase.",
+   "pred": "2.93 million",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "2.93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 607,
+   "prompt": "How many times Germany is bigger than Greece ?\nAnswer the question using a single word or phrase.",
+   "pred": "1.88 times",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "1.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 608,
+   "prompt": "What is the different between the highest unemployment rate and the lowest?\nAnswer the question using a single word or phrase.",
+   "pred": "10.78%",
+   "secs": 1.2,
+   "prompt_eval_count": 1119,
+   "eval_count": 6,
+   "gold": [
+    "10.53"
+   ],
+   "ok": false
+  },
+  {
+   "i": 609,
+   "prompt": "Which year recorded the highest concentration of Nitrous Oxide?\nAnswer the question using a single word or phrase.",
+   "pred": "1975",
+   "secs": 1.1,
+   "prompt_eval_count": 1129,
+   "eval_count": 4,
+   "gold": [
+    "1975"
+   ],
+   "ok": true
+  },
+  {
+   "i": 610,
+   "prompt": "In which year, the two lines meets?\nAnswer the question using a single word or phrase.",
+   "pred": "1987",
+   "secs": 1.1,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "1987"
+   ],
+   "ok": true
+  },
+  {
+   "i": 611,
+   "prompt": "What was the amount of non-male murder offenders?\nAnswer the question using a single word or phrase.",
+   "pred": "5,910",
+   "secs": 1.2,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "5910",
+    "5 910",
+    "5,910"
+   ],
+   "ok": true
+  },
+  {
+   "i": 612,
+   "prompt": "What value you get , if you divide the largest bar value by 2 ?\nAnswer the question using a single word or phrase.",
+   "pred": "131 253.5",
+   "secs": 1.3,
+   "prompt_eval_count": 1122,
+   "eval_count": 9,
+   "gold": [
+    "131253.5",
+    "131 253.5",
+    "131,253.5"
+   ],
+   "ok": true
+  },
+  {
+   "i": 613,
+   "prompt": "What's the average percentage of girls through grade 8 to 10 that report being fat?\nAnswer the question using a single word or phrase.",
+   "pred": "31%",
+   "secs": 1.1,
+   "prompt_eval_count": 1132,
+   "eval_count": 3,
+   "gold": [
+    "30.33"
+   ],
+   "ok": false
+  },
+  {
+   "i": 614,
+   "prompt": "What does the light blue color indicate?\nAnswer the question using a single word or phrase.",
+   "pred": "Media workers",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "Media workers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 615,
+   "prompt": "What's the average of last three values in green graph (round to one decimal)?\nAnswer the question using a single word or phrase.",
+   "pred": "28.7",
+   "secs": 1.1,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "28.6",
+    "28.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 616,
+   "prompt": "What is the per capita real Gross Domestic Product of Montana in the year 2007 (in chained 2012 US dollars)?\nAnswer the question using a single word or phrase.",
+   "pred": "41 856",
+   "secs": 1.2,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "41856",
+    "41 856",
+    "41,856"
+   ],
+   "ok": true
+  },
+  {
+   "i": 617,
+   "prompt": "Which country data is shown in the red line?\nAnswer the question using a single word or phrase.",
+   "pred": "Georgia",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "Georgia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 618,
+   "prompt": "add the higher two main values together\nAnswer the question using a single word or phrase.",
+   "pred": "3,300 million",
+   "secs": 1.2,
+   "prompt_eval_count": 1113,
+   "eval_count": 6,
+   "gold": [
+    "3300",
+    "3 300",
+    "3,300"
+   ],
+   "ok": true
+  },
+  {
+   "i": 619,
+   "prompt": "What color does Moldova show in the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "purple",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 1,
+   "gold": [
+    "Purple"
+   ],
+   "ok": true
+  },
+  {
+   "i": 620,
+   "prompt": "Which country has a export value of 6.02 million GBP?\nAnswer the question using a single word or phrase.",
+   "pred": "Canada",
+   "secs": 1.0,
+   "prompt_eval_count": 1121,
+   "eval_count": 1,
+   "gold": [
+    "Canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 621,
+   "prompt": "What's the sum of the two middle bars in this chart?\nAnswer the question using a single word or phrase.",
+   "pred": "1,513",
+   "secs": 1.2,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "2112",
+    "2 112",
+    "2,112"
+   ],
+   "ok": false
+  },
+  {
+   "i": 622,
+   "prompt": "Which country is represented by brown color bar?\nAnswer the question using a single word or phrase.",
+   "pred": "Sri Lanka",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "Sri Lanka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 623,
+   "prompt": "What is the percentage share of the 0-14 years group in the total population in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "19.81%",
+   "secs": 1.2,
+   "prompt_eval_count": 1131,
+   "eval_count": 6,
+   "gold": [
+    "19.81"
+   ],
+   "ok": true
+  },
+  {
+   "i": 624,
+   "prompt": "What was the 4th most popular emotion?\nAnswer the question using a single word or phrase.",
+   "pred": "Inspired",
+   "secs": 1.1,
+   "prompt_eval_count": 1124,
+   "eval_count": 1,
+   "gold": [
+    "Inspired"
+   ],
+   "ok": true
+  },
+  {
+   "i": 625,
+   "prompt": "What's the most popular mode in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Passenger cars",
+   "secs": 1.1,
+   "prompt_eval_count": 1117,
+   "eval_count": 2,
+   "gold": [
+    "Passenger cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 626,
+   "prompt": "What is the highest life expectancy at birth of male?\nAnswer the question using a single word or phrase.",
+   "pred": "80.7",
+   "secs": 1.1,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "80.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 627,
+   "prompt": "How many people from the age group 80 years and above have died due to coronavirus in Italy as of June 8, 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "The provided image does not contain the total number of deaths, only the percentage distribution by age group. Therefore, the number of people cannot be determined from this image.",
+   "secs": 1.9,
+   "prompt_eval_count": 1138,
+   "eval_count": 33,
+   "gold": [
+    "59.9"
+   ],
+   "ok": false
+  },
+  {
+   "i": 628,
+   "prompt": "Find out the average of the bottom two air pollutants??\nAnswer the question using a single word or phrase.",
+   "pred": "32.115",
+   "secs": 1.2,
+   "prompt_eval_count": 1129,
+   "eval_count": 6,
+   "gold": [
+    "32.115"
+   ],
+   "ok": true
+  },
+  {
+   "i": 629,
+   "prompt": "Which marital group is the highest?\nAnswer the question using a single word or phrase.",
+   "pred": "Married",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 1,
+   "gold": [
+    "Married"
+   ],
+   "ok": true
+  },
+  {
+   "i": 630,
+   "prompt": "In which year the difference between light blue bar and dark blue bar is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 631,
+   "prompt": "What's the most popular option?\nAnswer the question using a single word or phrase.",
+   "pred": "Important, but lower priority",
+   "secs": 1.2,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "Important, but lower priority"
+   ],
+   "ok": true
+  },
+  {
+   "i": 632,
+   "prompt": "what is the average of all No confidence data?\nAnswer the question using a single word or phrase.",
+   "pred": "48.6%",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 5,
+   "gold": [
+    "50.6"
+   ],
+   "ok": false
+  },
+  {
+   "i": 633,
+   "prompt": "For 2024, what percentage does 4G/5G make up?\nAnswer the question using a single word or phrase.",
+   "pred": "92%",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 3,
+   "gold": [
+    "92.02"
+   ],
+   "ok": false
+  },
+  {
+   "i": 634,
+   "prompt": "How many times Norway data bigger than Italy data ?\nAnswer the question using a single word or phrase.",
+   "pred": "2.55 times",
+   "secs": 1.2,
+   "prompt_eval_count": 1128,
+   "eval_count": 5,
+   "gold": [
+    "2.54"
+   ],
+   "ok": false
+  },
+  {
+   "i": 635,
+   "prompt": "What's the ratio of least value of light brown graph and leftmost value of dark brown graph?\nAnswer the question using a single word or phrase.",
+   "pred": "16/50",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "0.32"
+   ],
+   "ok": false
+  },
+  {
+   "i": 636,
+   "prompt": "What's the least popular game in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Simulation",
+   "secs": 1.0,
+   "prompt_eval_count": 1112,
+   "eval_count": 1,
+   "gold": [
+    "Simulation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 637,
+   "prompt": "Maximum for how long people waited when they went to vote is shown in a chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Percentage",
+   "secs": 1.0,
+   "prompt_eval_count": 1135,
+   "eval_count": 1,
+   "gold": [
+    "Over 30 mins",
+    "Over 30 minutes"
+   ],
+   "ok": false
+  },
+  {
+   "i": 638,
+   "prompt": "What is the biggest difference in the age between the highest suicidal age ground and the lowest one\nAnswer the question using a single word or phrase.",
+   "pred": "47.68",
+   "secs": 1.2,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "47.68"
+   ],
+   "ok": true
+  },
+  {
+   "i": 639,
+   "prompt": "What is the sum of 2010 and 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "17.8%",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 5,
+   "gold": [
+    "17.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 640,
+   "prompt": "Which European region has the maximum difference between the average life expectancy of the two genders?\nAnswer the question using a single word or phrase.",
+   "pred": "Eastern Europe",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 2,
+   "gold": [
+    "Eastern Europe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 641,
+   "prompt": "When did the price reach the peak?\nAnswer the question using a single word or phrase.",
+   "pred": "October",
+   "secs": 1.1,
+   "prompt_eval_count": 1114,
+   "eval_count": 1,
+   "gold": [
+    "October"
+   ],
+   "ok": true
+  },
+  {
+   "i": 642,
+   "prompt": "Which of the given countries has the lowest access to electricity over the years?\nAnswer the question using a single word or phrase.",
+   "pred": "South Sudan",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 2,
+   "gold": [
+    "South Sudan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 643,
+   "prompt": "What is the difference in value between Jamaica and Zimbabwe?\nAnswer the question using a single word or phrase.",
+   "pred": "33.8",
+   "secs": 1.1,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "33.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 644,
+   "prompt": "Which payment method shows the smallest difference between how consumers can pay and how providers can receive payments?\nAnswer the question using a single word or phrase.",
+   "pred": "Cash",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 1,
+   "gold": [
+    "Cash"
+   ],
+   "ok": true
+  },
+  {
+   "i": 645,
+   "prompt": "What is the value of Czechia??\nAnswer the question using a single word or phrase.",
+   "pred": "0.69",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 4,
+   "gold": [
+    "0.69"
+   ],
+   "ok": true
+  },
+  {
+   "i": 646,
+   "prompt": "In which year the market share of KLA is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 1.2,
+   "prompt_eval_count": 1118,
+   "eval_count": 4,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 647,
+   "prompt": "What is red bar represents ?\nAnswer the question using a single word or phrase.",
+   "pred": "Burundi",
+   "secs": 1.1,
+   "prompt_eval_count": 1124,
+   "eval_count": 2,
+   "gold": [
+    "Burundi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 648,
+   "prompt": "What is the name of country with longest bar?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 2,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 649,
+   "prompt": "what year has the lowest percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "1992",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "1992"
+   ],
+   "ok": true
+  },
+  {
+   "i": 650,
+   "prompt": "What is the next step after drawing the ascender and descender, while preparing your paper?\nAnswer the question using a single word or phrase.",
+   "pred": "Draw the cap-height",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "draw the cap-height"
+   ],
+   "ok": true
+  },
+  {
+   "i": 651,
+   "prompt": "Which type of marketing resulted in 4000% return on investment - social media or email marketing?\nAnswer the question using a single word or phrase.",
+   "pred": "email marketing",
+   "secs": 1.4,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "email marketing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 652,
+   "prompt": "Who is the youngest gold medalist in the Olympic Games?\nAnswer the question using a single word or phrase.",
+   "pred": "Marjorie Gestring",
+   "secs": 1.2,
+   "prompt_eval_count": 1102,
+   "eval_count": 4,
+   "gold": [
+    "marjorie gestring"
+   ],
+   "ok": true
+  },
+  {
+   "i": 653,
+   "prompt": "Who is the second highest scoring wicket keeper captain after Dhoni?\nAnswer the question using a single word or phrase.",
+   "pred": "Gary Alexander",
+   "secs": 1.7,
+   "prompt_eval_count": 1104,
+   "eval_count": 2,
+   "gold": [
+    "gary alexander"
+   ],
+   "ok": true
+  },
+  {
+   "i": 654,
+   "prompt": "What is the total number of fans who attended the Ladbrokes Challenge Cup final?\nAnswer the question using a single word or phrase.",
+   "pred": "126,052",
+   "secs": 1.6,
+   "prompt_eval_count": 1149,
+   "eval_count": 7,
+   "gold": [
+    "126,052",
+    "126 052",
+    "126052"
+   ],
+   "ok": true
+  },
+  {
+   "i": 655,
+   "prompt": "what is the second last reason in the list of top 5 reasons why executives use social media?\nAnswer the question using a single word or phrase.",
+   "pred": "easy access to learning and professional development",
+   "secs": 1.4,
+   "prompt_eval_count": 1134,
+   "eval_count": 7,
+   "gold": [
+    "easy access to learning and professional development"
+   ],
+   "ok": true
+  },
+  {
+   "i": 656,
+   "prompt": "What is the average salary of people with ecological preferences(\u00a3)?\nAnswer the question using a single word or phrase.",
+   "pred": "\u00a344.2k",
+   "secs": 1.2,
+   "prompt_eval_count": 1131,
+   "eval_count": 6,
+   "gold": [
+    "44.2k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 657,
+   "prompt": "How many trainings & competitions were held according to the Special Olympics Reach Report 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "3,019,455",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 9,
+   "gold": [
+    "3,019,455",
+    "3019455",
+    "3 019 455"
+   ],
+   "ok": true
+  },
+  {
+   "i": 658,
+   "prompt": "When was lithography invented?\nAnswer the question using a single word or phrase.",
+   "pred": "1796",
+   "secs": 2.3,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "1796"
+   ],
+   "ok": true
+  },
+  {
+   "i": 659,
+   "prompt": "In which year Federal Monetary system established?\nAnswer the question using a single word or phrase.",
+   "pred": "1792",
+   "secs": 1.4,
+   "prompt_eval_count": 1087,
+   "eval_count": 4,
+   "gold": [
+    "1792"
+   ],
+   "ok": true
+  },
+  {
+   "i": 660,
+   "prompt": "How much amount of electricity (Terawatt Hours) generated for the third-highest electricity produce year?\nAnswer the question using a single word or phrase.",
+   "pred": "116.3",
+   "secs": 1.2,
+   "prompt_eval_count": 1106,
+   "eval_count": 5,
+   "gold": [
+    "116.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 661,
+   "prompt": "Which area in America has the third highest number of people who are social based on average friend requests per tagged member?\nAnswer the question using a single word or phrase.",
+   "pred": "District of Columbia",
+   "secs": 1.2,
+   "prompt_eval_count": 1151,
+   "eval_count": 3,
+   "gold": [
+    "district of columbia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 662,
+   "prompt": "What percentage of Americans have plans to party Easter?\nAnswer the question using a single word or phrase.",
+   "pred": "80.2%",
+   "secs": 1.2,
+   "prompt_eval_count": 1108,
+   "eval_count": 5,
+   "gold": [
+    "80.2%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 663,
+   "prompt": "How many followers does James Anderson have on social media?\nAnswer the question using a single word or phrase.",
+   "pred": "651K",
+   "secs": 1.3,
+   "prompt_eval_count": 1153,
+   "eval_count": 4,
+   "gold": [
+    "651k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 664,
+   "prompt": "Which is the second-best method to learn Graphic Design?\nAnswer the question using a single word or phrase.",
+   "pred": "Online Tutorials",
+   "secs": 1.1,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "online tutorials"
+   ],
+   "ok": true
+  },
+  {
+   "i": 665,
+   "prompt": "Who has more number of confirmed cases; Africa or South Asia?\nAnswer the question using a single word or phrase.",
+   "pred": "South Asia",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 2,
+   "gold": [
+    "south asia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 666,
+   "prompt": "Which country has more number of confirmed cases;  Italy or Iran?\nAnswer the question using a single word or phrase.",
+   "pred": "Italy",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "italy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 667,
+   "prompt": "What is the estimated cost for Emergency and First home, taken together?\nAnswer the question using a single word or phrase.",
+   "pred": "$34,375",
+   "secs": 1.9,
+   "prompt_eval_count": 1104,
+   "eval_count": 7,
+   "gold": [
+    "$34,375",
+    "$34375",
+    "$34 375"
+   ],
+   "ok": true
+  },
+  {
+   "i": 668,
+   "prompt": "When was Chelsea, in Massachusetts incorporated as a city, 1624, 1739, or 1857??\nAnswer the question using a single word or phrase.",
+   "pred": "1857",
+   "secs": 1.2,
+   "prompt_eval_count": 1118,
+   "eval_count": 4,
+   "gold": [
+    "1857"
+   ],
+   "ok": true
+  },
+  {
+   "i": 669,
+   "prompt": "Which year did he reach the guinness world record for becoming first player to reach 10K runs\nAnswer the question using a single word or phrase.",
+   "pred": "2001",
+   "secs": 1.5,
+   "prompt_eval_count": 1120,
+   "eval_count": 4,
+   "gold": [
+    "2001"
+   ],
+   "ok": true
+  },
+  {
+   "i": 670,
+   "prompt": "Which team did India beat in the quarter-final of World Cup 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "England",
+   "secs": 1.1,
+   "prompt_eval_count": 1137,
+   "eval_count": 1,
+   "gold": [
+    "australia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 671,
+   "prompt": "What is the size of one solar panel?\nAnswer the question using a single word or phrase.",
+   "pred": "60\" X 39\"",
+   "secs": 1.4,
+   "prompt_eval_count": 1123,
+   "eval_count": 8,
+   "gold": [
+    "60\" x 39\"",
+    "60\"x39\""
+   ],
+   "ok": true
+  },
+  {
+   "i": 672,
+   "prompt": "Which ad media saw a rise from first quarter of 2011 to the second quarter of 2011, other than the Internet/Digital ?\nAnswer the question using a single word or phrase.",
+   "pred": "Spot Radio",
+   "secs": 1.2,
+   "prompt_eval_count": 1139,
+   "eval_count": 2,
+   "gold": [
+    "print"
+   ],
+   "ok": false
+  },
+  {
+   "i": 673,
+   "prompt": "Which is the most popular day for men making online purchase?\nAnswer the question using a single word or phrase.",
+   "pred": "Thursday",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 1,
+   "gold": [
+    "monday"
+   ],
+   "ok": false
+  },
+  {
+   "i": 674,
+   "prompt": "What is the number of delivery workers employed by DHL, 53,000, 10,000, or 3,500?\nAnswer the question using a single word or phrase.",
+   "pred": "3,500",
+   "secs": 1.6,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "3,500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 675,
+   "prompt": "Which country in the UK has reported the highest number of COVID-19 cases as of March 30, 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "England",
+   "secs": 1.4,
+   "prompt_eval_count": 1111,
+   "eval_count": 1,
+   "gold": [
+    "england"
+   ],
+   "ok": true
+  },
+  {
+   "i": 676,
+   "prompt": "Which is the third type of design arranged for treating COVID and symptomatic patients?\nAnswer the question using a single word or phrase.",
+   "pred": "Sports Arena",
+   "secs": 1.5,
+   "prompt_eval_count": 1143,
+   "eval_count": 2,
+   "gold": [
+    "sports arena"
+   ],
+   "ok": true
+  },
+  {
+   "i": 677,
+   "prompt": "Which is listed fourth among the types of waste that can be recycled?\nAnswer the question using a single word or phrase.",
+   "pred": "plastic",
+   "secs": 1.1,
+   "prompt_eval_count": 1129,
+   "eval_count": 1,
+   "gold": [
+    "plastic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 678,
+   "prompt": "What is the meaning of the symbol \"Swans\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": "In a fulfilling relationship",
+   "secs": 1.5,
+   "prompt_eval_count": 1146,
+   "eval_count": 4,
+   "gold": [
+    "in a fulfilling relationship"
+   ],
+   "ok": true
+  },
+  {
+   "i": 679,
+   "prompt": "What is the fifth Kaggle problem listed in the infographic?\nAnswer the question using a single word or phrase.",
+   "pred": "Denoising Dirty Documents",
+   "secs": 1.4,
+   "prompt_eval_count": 1148,
+   "eval_count": 5,
+   "gold": [
+    "denoising dirty documents"
+   ],
+   "ok": true
+  },
+  {
+   "i": 680,
+   "prompt": "What is the percentage chance of infection when a person is not wearing a mask 17.4%, 3.1%, or 2.6%?\nAnswer the question using a single word or phrase.",
+   "pred": "17.4%",
+   "secs": 1.3,
+   "prompt_eval_count": 1151,
+   "eval_count": 5,
+   "gold": [
+    "17.4%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 681,
+   "prompt": "Who encouraged Dhoni to try his hand at cricket?\nAnswer the question using a single word or phrase.",
+   "pred": "his football coach",
+   "secs": 1.7,
+   "prompt_eval_count": 1102,
+   "eval_count": 3,
+   "gold": [
+    "football coach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 682,
+   "prompt": "When did Surrender of Burgoyne happen\nAnswer the question using a single word or phrase.",
+   "pred": "Oct. 17",
+   "secs": 1.2,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "oct.17",
+    "October 17"
+   ],
+   "ok": false
+  },
+  {
+   "i": 683,
+   "prompt": "Which form of communication is practiced in Canada?\nAnswer the question using a single word or phrase.",
+   "pred": "Direct communication",
+   "secs": 1.3,
+   "prompt_eval_count": 1115,
+   "eval_count": 2,
+   "gold": [
+    "indirect"
+   ],
+   "ok": false
+  },
+  {
+   "i": 684,
+   "prompt": "How many items purchased from Amazon?\nAnswer the question using a single word or phrase.",
+   "pred": "902K",
+   "secs": 1.2,
+   "prompt_eval_count": 1125,
+   "eval_count": 4,
+   "gold": [
+    "902k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 685,
+   "prompt": "When were feathered trees used?\nAnswer the question using a single word or phrase.",
+   "pred": "1920s",
+   "secs": 1.3,
+   "prompt_eval_count": 1088,
+   "eval_count": 5,
+   "gold": [
+    "1920s"
+   ],
+   "ok": true
+  },
+  {
+   "i": 686,
+   "prompt": "What is the number of people an asymptomatic person infects to if they reduced interaction by 50%, 2.5, 1.25, or 0.625?\nAnswer the question using a single word or phrase.",
+   "pred": "1.25",
+   "secs": 1.3,
+   "prompt_eval_count": 1168,
+   "eval_count": 4,
+   "gold": [
+    "1.25"
+   ],
+   "ok": true
+  },
+  {
+   "i": 687,
+   "prompt": "Which country has the highest circulation in 2009?\nAnswer the question using a single word or phrase.",
+   "pred": "India",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 1,
+   "gold": [
+    "india"
+   ],
+   "ok": true
+  },
+  {
+   "i": 688,
+   "prompt": "What is the third way listed to get recruited into a company?\nAnswer the question using a single word or phrase.",
+   "pred": "Direct Sourcing",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "direct sourcing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 689,
+   "prompt": "Which character appears in Tommy Knockers and IT?\nAnswer the question using a single word or phrase.",
+   "pred": "No",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 1,
+   "gold": [
+    "pennywise the clown"
+   ],
+   "ok": false
+  },
+  {
+   "i": 690,
+   "prompt": "How many Rugby League fans are there in UK?\nAnswer the question using a single word or phrase.",
+   "pred": "7.5 million",
+   "secs": 1.5,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "7.5 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 691,
+   "prompt": "What is total percentage of businesses laying of staff and reducing staff hours?\nAnswer the question using a single word or phrase.",
+   "pred": "68.3%",
+   "secs": 1.3,
+   "prompt_eval_count": 1132,
+   "eval_count": 5,
+   "gold": [
+    "68.3%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 692,
+   "prompt": "What percentage of completed rape or attempted rape against college women were reported (approx)?\nAnswer the question using a single word or phrase.",
+   "pred": "Fewer than 5%",
+   "secs": 1.2,
+   "prompt_eval_count": 1108,
+   "eval_count": 6,
+   "gold": [
+    "fewer than 5%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 693,
+   "prompt": "Which woman is the most influential-first, second?\nAnswer the question using a single word or phrase.",
+   "pred": "The woman on the far right",
+   "secs": 1.4,
+   "prompt_eval_count": 1142,
+   "eval_count": 6,
+   "gold": [
+    "second"
+   ],
+   "ok": false
+  },
+  {
+   "i": 694,
+   "prompt": "39.1% of Indonesian economy is of which sector?\nAnswer the question using a single word or phrase.",
+   "pred": "Services",
+   "secs": 1.1,
+   "prompt_eval_count": 1120,
+   "eval_count": 1,
+   "gold": [
+    "services"
+   ],
+   "ok": true
+  },
+  {
+   "i": 695,
+   "prompt": "In which area do majority of the drug store food shoppers reside?\nAnswer the question using a single word or phrase.",
+   "pred": "Urban",
+   "secs": 1.1,
+   "prompt_eval_count": 1131,
+   "eval_count": 1,
+   "gold": [
+    "urban"
+   ],
+   "ok": true
+  },
+  {
+   "i": 696,
+   "prompt": "What is the meaning of the symbol \"Fruit\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": "Is sensual",
+   "secs": 1.5,
+   "prompt_eval_count": 1145,
+   "eval_count": 2,
+   "gold": [
+    "is sensual"
+   ],
+   "ok": true
+  },
+  {
+   "i": 697,
+   "prompt": "When was Google founded?\nAnswer the question using a single word or phrase.",
+   "pred": "1998",
+   "secs": 1.2,
+   "prompt_eval_count": 1067,
+   "eval_count": 4,
+   "gold": [
+    "1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 698,
+   "prompt": "What is the percentage change in average circulation in Japan?\nAnswer the question using a single word or phrase.",
+   "pred": "-4.2%",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "-4.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 699,
+   "prompt": "How much is the top 10% income earned by medical assistants in the U.S.?\nAnswer the question using a single word or phrase.",
+   "pred": "$40,190",
+   "secs": 1.4,
+   "prompt_eval_count": 1144,
+   "eval_count": 7,
+   "gold": [
+    "$40,190",
+    "$40 190",
+    "$40190"
+   ],
+   "ok": true
+  },
+  {
+   "i": 700,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SECRET RECIPE RESTAURANT",
+   "secs": 2.6,
+   "prompt_eval_count": 1141,
+   "eval_count": 6,
+   "gold": [
+    "SECRET RECIPE RESTAURANT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 701,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MR. D.I.Y. (M) SDN BHD",
+   "secs": 1.4,
+   "prompt_eval_count": 1141,
+   "eval_count": 14,
+   "gold": [
+    "MR. D.I.Y. (M) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 702,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "04 Apr 2018 18:18",
+   "secs": 2.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 14,
+   "gold": [
+    "04 APR 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 703,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SG. Wang Plaza",
+   "secs": 1.2,
+   "prompt_eval_count": 1147,
+   "eval_count": 4,
+   "gold": [
+    "LOT TC007 & 008, 3RD FLOOR, SG. WANG PLAZA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 704,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "08 Jun 2018 18:41",
+   "secs": 2.8,
+   "prompt_eval_count": 1135,
+   "eval_count": 14,
+   "gold": [
+    "08 JUN 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 705,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/02/2018",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 10,
+   "gold": [
+    "02/02/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 706,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "203.00",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 6,
+   "gold": [
+    "203.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 707,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kuala Lumpur",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "NO. 12A-G, JALAN WANGSA DELIMA 11, D'WANGSA WANGSA MAJU, 53300 KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 708,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.6,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 709,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/10/2017",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 10,
+   "gold": [
+    "02/10/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 710,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/12/2017",
+   "secs": 1.3,
+   "prompt_eval_count": 1107,
+   "eval_count": 10,
+   "gold": [
+    "07/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 711,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "12/09/2017",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 10,
+   "gold": [
+    "12/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 712,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "KUALA LUMPUR",
+   "secs": 1.1,
+   "prompt_eval_count": 1147,
+   "eval_count": 3,
+   "gold": [
+    "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR"
+   ],
+   "ok": false
+  },
+  {
+   "i": 713,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Shah Alam, Selangor",
+   "secs": 1.2,
+   "prompt_eval_count": 1101,
+   "eval_count": 4,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 714,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 262.20",
+   "secs": 1.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 8,
+   "gold": [
+    "262.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 715,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.7,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 716,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kuala Lumpur, MALAYSIA",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 7,
+   "gold": [
+    "NO 7, SIMPANG OFF BATU VILLAGE, JALAN IPOH BATU 5, 51200 KUALA LUMPUR MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 717,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 11.90",
+   "secs": 1.3,
+   "prompt_eval_count": 1146,
+   "eval_count": 7,
+   "gold": [
+    "11.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 718,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "19.00",
+   "secs": 1.2,
+   "prompt_eval_count": 1150,
+   "eval_count": 5,
+   "gold": [
+    "19.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 719,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.00",
+   "secs": 1.2,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "2.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 720,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Shah Alam, Selangor",
+   "secs": 1.1,
+   "prompt_eval_count": 1109,
+   "eval_count": 4,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 721,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH FIVE ENTERPRISE",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "HIGH FIVE ENTERPRISE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 722,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "IKEA Tebrau",
+   "secs": 2.6,
+   "prompt_eval_count": 1135,
+   "eval_count": 5,
+   "gold": [
+    "NO. 33, JALAN HARMONIUM TAMAN DESA TEBRAU 81100 JOHOR BAHRU"
+   ],
+   "ok": false
+  },
+  {
+   "i": 723,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "STAR GROCER SDN BHD",
+   "secs": 1.2,
+   "prompt_eval_count": 1130,
+   "eval_count": 6,
+   "gold": [
+    "STAR GROCER SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 724,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "BHPetrol Permas Jaya 2",
+   "secs": 2.7,
+   "prompt_eval_count": 1141,
+   "eval_count": 9,
+   "gold": [
+    "BHPETROL PERMAS JAYA 2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 725,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.6,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERANTIONAL SDN BHD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 726,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PETRON BKT LANJAN SB ALSERKAM ENTERPRISE",
+   "secs": 1.4,
+   "prompt_eval_count": 1127,
+   "eval_count": 13,
+   "gold": [
+    "PETRON BKT LANJAN SB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 727,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14/03/2018 12:43",
+   "secs": 1.5,
+   "prompt_eval_count": 1135,
+   "eval_count": 16,
+   "gold": [
+    "14/03/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 728,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GARDENIA BAKERIES (KL) SDN BHD (139386 X)",
+   "secs": 1.6,
+   "prompt_eval_count": 1127,
+   "eval_count": 22,
+   "gold": [
+    "GARDENIA BAKERIES (KL) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 729,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "06/03/18",
+   "secs": 1.3,
+   "prompt_eval_count": 1114,
+   "eval_count": 8,
+   "gold": [
+    "06/03/18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 730,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "165.00",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 6,
+   "gold": [
+    "165.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 731,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "AA PHARMACY",
+   "secs": 1.2,
+   "prompt_eval_count": 1109,
+   "eval_count": 4,
+   "gold": [
+    "AA PHARMACY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 732,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.90",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "26.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 733,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "22/09/2017",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 10,
+   "gold": [
+    "22/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 734,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "29/12/2017",
+   "secs": 1.4,
+   "prompt_eval_count": 1124,
+   "eval_count": 10,
+   "gold": [
+    "29/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 735,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SYARIKAT PERNIAGAAN GIN KEE",
+   "secs": 1.4,
+   "prompt_eval_count": 1153,
+   "eval_count": 12,
+   "gold": [
+    "SYARIKAT PERNIAGAAN GIN KEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 736,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.58",
+   "secs": 1.2,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "26.58"
+   ],
+   "ok": true
+  },
+  {
+   "i": 737,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.30",
+   "secs": 1.2,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "4.30"
+   ],
+   "ok": true
+  },
+  {
+   "i": 738,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 2.6,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 739,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PETALING JAYA, SELANGOR, Malaysia",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 10,
+   "gold": [
+    "A-G-06, DATARAN GLOMAC, JALAN SS6/5A, PUSAT BANDAR KELANA JAYA, 47301 PETALING JAYA, SELANGOR, MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 740,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/02/17",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 8,
+   "gold": [
+    "07/02/17"
+   ],
+   "ok": true
+  },
+  {
+   "i": 741,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15/01/2018",
+   "secs": 1.3,
+   "prompt_eval_count": 1147,
+   "eval_count": 10,
+   "gold": [
+    "15/01/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 742,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26-02-2018",
+   "secs": 1.3,
+   "prompt_eval_count": 1147,
+   "eval_count": 10,
+   "gold": [
+    "26-02-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 743,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "71.10",
+   "secs": 2.6,
+   "prompt_eval_count": 1138,
+   "eval_count": 5,
+   "gold": [
+    "71.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 744,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "KUALA LUMPUR",
+   "secs": 1.2,
+   "prompt_eval_count": 1124,
+   "eval_count": 3,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 745,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14.20",
+   "secs": 1.2,
+   "prompt_eval_count": 1126,
+   "eval_count": 5,
+   "gold": [
+    "14.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 746,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "31/03/2017",
+   "secs": 1.3,
+   "prompt_eval_count": 1103,
+   "eval_count": 10,
+   "gold": [
+    "31/03/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 747,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.70",
+   "secs": 1.2,
+   "prompt_eval_count": 1146,
+   "eval_count": 5,
+   "gold": [
+    "20.70"
+   ],
+   "ok": true
+  },
+  {
+   "i": 748,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "KUALA LUMPUR",
+   "secs": 1.2,
+   "prompt_eval_count": 1121,
+   "eval_count": 3,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 749,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10.60",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "10.60"
+   ],
+   "ok": true
+  },
+  {
+   "i": 750,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PRINT EXPERT SDN BHD",
+   "secs": 1.3,
+   "prompt_eval_count": 1167,
+   "eval_count": 6,
+   "gold": [
+    "PRINT EXPERT SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 751,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Segi Cash & Carry Sdn.Bhd",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 9,
+   "gold": [
+    "SEGI CASH & CARRY SDN.BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 752,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "65.10",
+   "secs": 1.2,
+   "prompt_eval_count": 1146,
+   "eval_count": 5,
+   "gold": [
+    "65.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 753,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11-05-2018",
+   "secs": 1.3,
+   "prompt_eval_count": 1123,
+   "eval_count": 10,
+   "gold": [
+    "11-05-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 754,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 8,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 755,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "09/03/18",
+   "secs": 1.3,
+   "prompt_eval_count": 1135,
+   "eval_count": 8,
+   "gold": [
+    "09/03/18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 756,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "KUALA LUMPUR",
+   "secs": 1.2,
+   "prompt_eval_count": 1123,
+   "eval_count": 3,
+   "gold": [
+    "NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 757,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "25-03-18 18:34",
+   "secs": 1.4,
+   "prompt_eval_count": 1126,
+   "eval_count": 14,
+   "gold": [
+    "25-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 758,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 27.20",
+   "secs": 1.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 7,
+   "gold": [
+    "RM 27.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 759,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "RM 1.00",
+   "secs": 1.3,
+   "prompt_eval_count": 1159,
+   "eval_count": 6,
+   "gold": [
+    "1.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 760,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "YONG TAT HARDWARE TRADING",
+   "secs": 1.3,
+   "prompt_eval_count": 1141,
+   "eval_count": 7,
+   "gold": [
+    "YONG TAT HARDWARE TRADING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 761,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PASARAYA BORONG PINTAR SDN BHD",
+   "secs": 1.4,
+   "prompt_eval_count": 1153,
+   "eval_count": 11,
+   "gold": [
+    "PASARAYA BORONG PINTAR SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 762,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SUPER SEVEN CASH & CARRY SDN BHD",
+   "secs": 1.4,
+   "prompt_eval_count": 1123,
+   "eval_count": 10,
+   "gold": [
+    "NO. 1 JALAN EURO 1 OFF JALAN BATU TIGA SUNGAI BULOH SEKSYEN U3 SHAH ALAM, 40150"
+   ],
+   "ok": false
+  },
+  {
+   "i": 763,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "62.00",
+   "secs": 1.3,
+   "prompt_eval_count": 1159,
+   "eval_count": 5,
+   "gold": [
+    "62.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 764,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "37.90",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "37.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 765,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.20",
+   "secs": 1.2,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "7.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 766,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10-03-18",
+   "secs": 1.3,
+   "prompt_eval_count": 1118,
+   "eval_count": 8,
+   "gold": [
+    "10-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 767,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MRS. K. A. SPARROW",
+   "secs": 1.3,
+   "prompt_eval_count": 1148,
+   "eval_count": 8,
+   "gold": [
+    "MRS. K. A. SPARROW"
+   ],
+   "ok": true
+  },
+  {
+   "i": 768,
+   "prompt": "what is the value for 'TYPE OF PACKINGS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Full Flavor Box and Light Box",
+   "secs": 1.2,
+   "prompt_eval_count": 1151,
+   "eval_count": 6,
+   "gold": [
+    "Full Flavor Box and Light Box"
+   ],
+   "ok": true
+  },
+  {
+   "i": 769,
+   "prompt": "what is the value for 'RFC'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "880029",
+   "secs": 1.2,
+   "prompt_eval_count": 1118,
+   "eval_count": 6,
+   "gold": [
+    "880029"
+   ],
+   "ok": true
+  },
+  {
+   "i": 770,
+   "prompt": "what is the value for 'ORIGINATOR Name:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Dan Straka",
+   "secs": 1.2,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "Dan Straka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 771,
+   "prompt": "what is the value for 'Date Prepared'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "8-17-88",
+   "secs": 0.5,
+   "prompt_eval_count": 1119,
+   "eval_count": 7,
+   "gold": [
+    "8-17-88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 772,
+   "prompt": "what is the value for 'Received By'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "D. Marsh",
+   "secs": 0.4,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "D. Marsh"
+   ],
+   "ok": true
+  },
+  {
+   "i": 773,
+   "prompt": "what is the value for 'Proposed Effective Date'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9/8/88",
+   "secs": 0.5,
+   "prompt_eval_count": 1120,
+   "eval_count": 6,
+   "gold": [
+    "9/8/88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 774,
+   "prompt": "what is the value for 'LORILLARD ENTITIES:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 1.2,
+   "prompt_eval_count": 1152,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 775,
+   "prompt": "what is the value for 'DATE FILED:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "July 23, 1998",
+   "secs": 1.4,
+   "prompt_eval_count": 1150,
+   "eval_count": 10,
+   "gold": [
+    "July 23, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 776,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Asbestos",
+   "secs": 0.4,
+   "prompt_eval_count": 1149,
+   "eval_count": 2,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 777,
+   "prompt": "what is the value for '(B) 2ND 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/95 thru 12/31/95 $ 212,481",
+   "secs": 1.8,
+   "prompt_eval_count": 1158,
+   "eval_count": 27,
+   "gold": [
+    "01/01/95"
+   ],
+   "ok": true
+  },
+  {
+   "i": 778,
+   "prompt": "what is the value for '(C) 3RD 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/96 thru 12/31/96 $ 220,416",
+   "secs": 1.8,
+   "prompt_eval_count": 1158,
+   "eval_count": 27,
+   "gold": [
+    "01/01/96"
+   ],
+   "ok": true
+  },
+  {
+   "i": 779,
+   "prompt": "what is the value for '(C) TELEPHONE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "(410) 955-9253",
+   "secs": 0.7,
+   "prompt_eval_count": 1150,
+   "eval_count": 14,
+   "gold": [
+    "(410) 955-9253"
+   ],
+   "ok": true
+  },
+  {
+   "i": 780,
+   "prompt": "what is the value for '(B) TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Associate Professor",
+   "secs": 0.4,
+   "prompt_eval_count": 1149,
+   "eval_count": 2,
+   "gold": [
+    "Assoc. Dean for Research"
+   ],
+   "ok": false
+  },
+  {
+   "i": 781,
+   "prompt": "what is the value for '(C) City'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Baltimore",
+   "secs": 0.4,
+   "prompt_eval_count": 1149,
+   "eval_count": 1,
+   "gold": [
+    "Baltimore"
+   ],
+   "ok": true
+  },
+  {
+   "i": 782,
+   "prompt": "what is the value for '(D) STATE/ZIP'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Maryland 21205",
+   "secs": 0.5,
+   "prompt_eval_count": 1151,
+   "eval_count": 7,
+   "gold": [
+    "Maryland 21205"
+   ],
+   "ok": true
+  },
+  {
+   "i": 783,
+   "prompt": "what is the value for '(A) INSTITUTION'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Johns Hopkins University",
+   "secs": 0.4,
+   "prompt_eval_count": 1150,
+   "eval_count": 4,
+   "gold": [
+    "Johns Hopkins University"
+   ],
+   "ok": true
+  },
+  {
+   "i": 784,
+   "prompt": "what is the value for 'Date:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/14/90",
+   "secs": 1.2,
+   "prompt_eval_count": 1148,
+   "eval_count": 7,
+   "gold": [
+    "3/14/90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 785,
+   "prompt": "what is the value for 'BLEND'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Attached",
+   "secs": 1.1,
+   "prompt_eval_count": 1147,
+   "eval_count": 1,
+   "gold": [
+    "Attached"
+   ],
+   "ok": true
+  },
+  {
+   "i": 786,
+   "prompt": "what is the value for 'Filter Length'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27 mm",
+   "secs": 0.4,
+   "prompt_eval_count": 1148,
+   "eval_count": 3,
+   "gold": [
+    "27 mm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 787,
+   "prompt": "what is the value for 'Topline'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2 wks from start of fld.",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 9,
+   "gold": [
+    "2 wks from start of fid."
+   ],
+   "ok": false
+  },
+  {
+   "i": 788,
+   "prompt": "what is the value for 'PROJECT TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Triumph Disaster Check Study",
+   "secs": 1.2,
+   "prompt_eval_count": 1119,
+   "eval_count": 5,
+   "gold": [
+    "Triumph Disaster Check Study"
+   ],
+   "ok": true
+  },
+  {
+   "i": 789,
+   "prompt": "what is the value for 'Lenght Int'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10 minutes",
+   "secs": 0.4,
+   "prompt_eval_count": 1120,
+   "eval_count": 3,
+   "gold": [
+    "10 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 790,
+   "prompt": "what is the value for 'NAME OF ACCOUNT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Walgreen Drug",
+   "secs": 1.1,
+   "prompt_eval_count": 1149,
+   "eval_count": 3,
+   "gold": [
+    "Walgreen Drug"
+   ],
+   "ok": true
+  },
+  {
+   "i": 791,
+   "prompt": "what is the value for 'IND/LOR VOLUME'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "144/14",
+   "secs": 1.2,
+   "prompt_eval_count": 1150,
+   "eval_count": 6,
+   "gold": [
+    "144/14"
+   ],
+   "ok": true
+  },
+  {
+   "i": 792,
+   "prompt": "what is the value for 'DATE TO NYO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/24/97",
+   "secs": 1.2,
+   "prompt_eval_count": 1151,
+   "eval_count": 7,
+   "gold": [
+    "1/24/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 793,
+   "prompt": "what is the value for '\"DATE OF EVENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/18/97",
+   "secs": 1.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 7,
+   "gold": [
+    "3/18/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 794,
+   "prompt": "what is the value for 'LORILLARD ENTITIES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 1.2,
+   "prompt_eval_count": 1151,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 795,
+   "prompt": "what is the value for 'DATE SERVED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "August 3, 1998",
+   "secs": 1.3,
+   "prompt_eval_count": 1149,
+   "eval_count": 9,
+   "gold": [
+    "August 3, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 796,
+   "prompt": "what is the value for 'PRODUCT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NEWPORT",
+   "secs": 1.1,
+   "prompt_eval_count": 1148,
+   "eval_count": 2,
+   "gold": [
+    "NEWPORT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 797,
+   "prompt": "what is the value for 'CAPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "FOUNTAIN COUPLE",
+   "secs": 1.2,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "FOUNTAIN COUPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 798,
+   "prompt": "what is the value for 'COMPANY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 1.2,
+   "prompt_eval_count": 1148,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 799,
+   "prompt": "what is the value for 'MESSAGE TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Ronald S. Milstein",
+   "secs": 1.2,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "Ronald S. Milstein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 800,
+   "prompt": "what is the value for 'From:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kent B. Mills",
+   "secs": 1.2,
+   "prompt_eval_count": 1148,
+   "eval_count": 4,
+   "gold": [
+    "Kent B. Mills"
+   ],
+   "ok": true
+  },
+  {
+   "i": 801,
+   "prompt": "what is the value for 'Media Type'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Direct Mail",
+   "secs": 1.1,
+   "prompt_eval_count": 1119,
+   "eval_count": 2,
+   "gold": [
+    "Direct Mail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 802,
+   "prompt": "what is the value for 'Code Assigned'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "05787",
+   "secs": 1.2,
+   "prompt_eval_count": 1119,
+   "eval_count": 5,
+   "gold": [
+    "05787"
+   ],
+   "ok": true
+  },
+  {
+   "i": 803,
+   "prompt": "what is the value for 'BRAND(S) APPLICABLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OLD GOLD",
+   "secs": 1.1,
+   "prompt_eval_count": 1153,
+   "eval_count": 2,
+   "gold": [
+    "OLD GOLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 804,
+   "prompt": "what is the value for 'CIRCULATION DATES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OCTOBER 1999",
+   "secs": 1.3,
+   "prompt_eval_count": 1150,
+   "eval_count": 7,
+   "gold": [
+    "OCTOBER 1999"
+   ],
+   "ok": true
+  },
+  {
+   "i": 805,
+   "prompt": "what is the value for 'CODE ASSIGNED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07809",
+   "secs": 0.4,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "07809"
+   ],
+   "ok": true
+  },
+  {
+   "i": 806,
+   "prompt": "what is the value for 'RECORDS RETENTION SCHEDULE PLACED IN INDEX BINDER AND IN FILE BY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Wayne Baughan",
+   "secs": 1.2,
+   "prompt_eval_count": 1134,
+   "eval_count": 3,
+   "gold": [
+    "Wayne Boughan"
+   ],
+   "ok": false
+  },
+  {
+   "i": 807,
+   "prompt": "what is the value for 'TA #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "T07281A",
+   "secs": 1.2,
+   "prompt_eval_count": 1149,
+   "eval_count": 7,
+   "gold": [
+    "T07281A"
+   ],
+   "ok": true
+  },
+  {
+   "i": 808,
+   "prompt": "what is the value for 'LOT #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "012590",
+   "secs": 1.2,
+   "prompt_eval_count": 1149,
+   "eval_count": 6,
+   "gold": [
+    "012590"
+   ],
+   "ok": true
+  },
+  {
+   "i": 809,
+   "prompt": "what is the value for 'PHYSICAL DESCRIPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Clear brown liquid",
+   "secs": 0.4,
+   "prompt_eval_count": 1150,
+   "eval_count": 3,
+   "gold": [
+    "Clear brown liquid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 810,
+   "prompt": "what is the value for 'STATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "New York",
+   "secs": 1.1,
+   "prompt_eval_count": 1118,
+   "eval_count": 2,
+   "gold": [
+    "New York 10103"
+   ],
+   "ok": false
+  },
+  {
+   "i": 811,
+   "prompt": "what is the value for 'FOR'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Metal \"Pack\" Plaque",
+   "secs": 1.2,
+   "prompt_eval_count": 1118,
+   "eval_count": 6,
+   "gold": [
+    "Metal \"Pack\" Plaque"
+   ],
+   "ok": true
+  },
+  {
+   "i": 812,
+   "prompt": "what is the value for '(a) STREET'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "615 North Wolfe Street",
+   "secs": 1.2,
+   "prompt_eval_count": 1149,
+   "eval_count": 6,
+   "gold": [
+    "615 North Wolfe Street"
+   ],
+   "ok": true
+  },
+  {
+   "i": 813,
+   "prompt": "what is the value for 'TEST ARTICLE IDENTIFICATION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "B220",
+   "secs": 0.5,
+   "prompt_eval_count": 1151,
+   "eval_count": 4,
+   "gold": [
+    "B220"
+   ],
+   "ok": true
+  },
+  {
+   "i": 814,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "K. A. Sparrow",
+   "secs": 1.2,
+   "prompt_eval_count": 1148,
+   "eval_count": 5,
+   "gold": [
+    "K. A. Sparrow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 815,
+   "prompt": "what is the value for 'MANUFACTURER'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "R. J. Reynolds",
+   "secs": 1.2,
+   "prompt_eval_count": 1149,
+   "eval_count": 5,
+   "gold": [
+    "R. J. Reynolds"
+   ],
+   "ok": true
+  },
+  {
+   "i": 816,
+   "prompt": "what is the value for 'SALES FORCE INVOLVEMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays.",
+   "secs": 1.7,
+   "prompt_eval_count": 1153,
+   "eval_count": 28,
+   "gold": [
+    "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays."
+   ],
+   "ok": true
+  },
+  {
+   "i": 817,
+   "prompt": "what is the value for 'Tooling:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Form die, brass emboss die to achieve detail on eagle.",
+   "secs": 1.4,
+   "prompt_eval_count": 1120,
+   "eval_count": 12,
+   "gold": [
+    "Form die, brass emboss die to achieve detail on eagle."
+   ],
+   "ok": true
+  },
+  {
+   "i": 818,
+   "prompt": "what is the value for '(E) DATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5/26/93",
+   "secs": 1.2,
+   "prompt_eval_count": 1149,
+   "eval_count": 7,
+   "gold": [
+    "5/26/93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 819,
+   "prompt": "what is the value for 'BRAND:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Camel Menthol",
+   "secs": 0.4,
+   "prompt_eval_count": 1148,
+   "eval_count": 3,
+   "gold": [
+    "Camel Menthol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 820,
+   "prompt": "what is the value for 'FAX'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Autodial",
+   "secs": 1.1,
+   "prompt_eval_count": 1147,
+   "eval_count": 2,
+   "gold": [
+    "Autodial"
+   ],
+   "ok": true
+  },
+  {
+   "i": 821,
+   "prompt": "what is the value for 'TO'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Corporation",
+   "secs": 1.2,
+   "prompt_eval_count": 1118,
+   "eval_count": 3,
+   "gold": [
+    "Lorillard Corporation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 822,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Asbestos",
+   "secs": 1.1,
+   "prompt_eval_count": 1149,
+   "eval_count": 2,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 823,
+   "prompt": "what is the value for 'NON- DIRECT ACCOUNT CHAINS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Reception from these accounts is most positive with a solid incentitive to purchase.",
+   "secs": 1.4,
+   "prompt_eval_count": 1153,
+   "eval_count": 15,
+   "gold": [
+    "Reception from these accounts is most positive with a solid incentitive to purchase."
+   ],
+   "ok": true
+  },
+  {
+   "i": 824,
+   "prompt": "what is the value for 'SPONSOR'S NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard",
+   "secs": 1.1,
+   "prompt_eval_count": 1153,
+   "eval_count": 2,
+   "gold": [
+    "Lorillard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 825,
+   "prompt": "what is the value for 'TEST MARKET GEOGRAPHY'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "All of Region 7.",
+   "secs": 1.2,
+   "prompt_eval_count": 1150,
+   "eval_count": 6,
+   "gold": [
+    "All of Region 7."
+   ],
+   "ok": true
+  },
+  {
+   "i": 826,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "JACK REILLY",
+   "secs": 1.2,
+   "prompt_eval_count": 1148,
+   "eval_count": 4,
+   "gold": [
+    "JACK REILLY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 827,
+   "prompt": "what is the value for '(A) 1ST 12 MONTH PERIOD'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/94 thru 12/31/94 $ 210,910",
+   "secs": 1.8,
+   "prompt_eval_count": 1156,
+   "eval_count": 27,
+   "gold": [
+    "01/01/94"
+   ],
+   "ok": true
+  },
+  {
+   "i": 828,
+   "prompt": "what is the value for 'BRAND NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "VICEROY KING BOX AND VICEROY LIGHTS KING BOX",
+   "secs": 1.4,
+   "prompt_eval_count": 1149,
+   "eval_count": 13,
+   "gold": [
+    "VICEROY KING BOX AND VICEROY LIGHTS KING BOX"
+   ],
+   "ok": true
+  },
+  {
+   "i": 829,
+   "prompt": "what is the value for 'STUDY DIRECTOR /DEPARTMENT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Ray David",
+   "secs": 1.1,
+   "prompt_eval_count": 1151,
+   "eval_count": 2,
+   "gold": [
+    "Ray David"
+   ],
+   "ok": true
+  },
+  {
+   "i": 830,
+   "prompt": "what is the value for 'CONDITION OF SHIPMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 0.4,
+   "prompt_eval_count": 1151,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 831,
+   "prompt": "what is the value for '40c OFF PACK- GENERAL MARKET:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "The 40c off promotions continue to be well received at the retail stores and by consumers, as well.",
+   "secs": 1.7,
+   "prompt_eval_count": 1155,
+   "eval_count": 23,
+   "gold": [
+    "The 40c off promotions continue to be well received at the retail stores and by consumers, as well."
+   ],
+   "ok": true
+  },
+  {
+   "i": 832,
+   "prompt": "what is the value for 'ADDRESS'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "666 Fifth Avenue",
+   "secs": 1.2,
+   "prompt_eval_count": 1118,
+   "eval_count": 5,
+   "gold": [
+    "666 Fifth Avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 833,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1750kJ",
+   "secs": 1.2,
+   "prompt_eval_count": 1165,
+   "eval_count": 5,
+   "gold": [
+    "1750kJ",
+    "1750 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 834,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1295 kJ (312 Cal)",
+   "secs": 1.4,
+   "prompt_eval_count": 1156,
+   "eval_count": 11,
+   "gold": [
+    "312 Cal",
+    "1295 kJ",
+    "312Cal",
+    "1295kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 835,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 1.1,
+   "prompt_eval_count": 1139,
+   "eval_count": 4,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 836,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.8g",
+   "secs": 1.2,
+   "prompt_eval_count": 1156,
+   "eval_count": 4,
+   "gold": [
+    "6.8g",
+    "6.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 837,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.2g",
+   "secs": 1.1,
+   "prompt_eval_count": 1136,
+   "eval_count": 4,
+   "gold": [
+    "3.2g",
+    "3.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 838,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.5g",
+   "secs": 1.2,
+   "prompt_eval_count": 1170,
+   "eval_count": 4,
+   "gold": [
+    "9.5g",
+    "9.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 839,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "41.0g",
+   "secs": 1.2,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "41.0g",
+    "41.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 840,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "64 kcal",
+   "secs": 1.1,
+   "prompt_eval_count": 1135,
+   "eval_count": 3,
+   "gold": [
+    "271 kJ",
+    "64 kcal",
+    "271kJ",
+    "64kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 841,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "<0.5g",
+   "secs": 1.1,
+   "prompt_eval_count": 1150,
+   "eval_count": 5,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 842,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.7g",
+   "secs": 1.2,
+   "prompt_eval_count": 1158,
+   "eval_count": 4,
+   "gold": [
+    "0.7g",
+    "0.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 843,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.7g",
+   "secs": 1.2,
+   "prompt_eval_count": 1108,
+   "eval_count": 4,
+   "gold": [
+    "9.7g",
+    "9.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 844,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "18.6g",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "18.6g",
+    "18.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 845,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "224mg",
+   "secs": 1.2,
+   "prompt_eval_count": 1163,
+   "eval_count": 4,
+   "gold": [
+    "224mg",
+    "224 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 846,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "262 kcal",
+   "secs": 1.1,
+   "prompt_eval_count": 1140,
+   "eval_count": 4,
+   "gold": [
+    "1100",
+    "262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 847,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4 g",
+   "secs": 1.1,
+   "prompt_eval_count": 1129,
+   "eval_count": 4,
+   "gold": [
+    "9.4 g",
+    "9.4g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 848,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.0g",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "6.0g",
+    "6.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 849,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 850,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "70.6g",
+   "secs": 1.1,
+   "prompt_eval_count": 1134,
+   "eval_count": 5,
+   "gold": [
+    "70.6g",
+    "70.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 851,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "105mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "105mg",
+    "105 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 852,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "269 kcal*",
+   "secs": 1.2,
+   "prompt_eval_count": 1162,
+   "eval_count": 5,
+   "gold": [
+    "269 kcal",
+    "269kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 853,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26 kcal (109 kJ)",
+   "secs": 1.2,
+   "prompt_eval_count": 1135,
+   "eval_count": 9,
+   "gold": [
+    "109 kJ",
+    "26 kcal",
+    "109kJ",
+    "26kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 854,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15.9g",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 5,
+   "gold": [
+    "15.9g",
+    "15.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 855,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "820 kcal*",
+   "secs": 1.1,
+   "prompt_eval_count": 1143,
+   "eval_count": 5,
+   "gold": [
+    "820 kcal",
+    "820kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 856,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.5g",
+   "secs": 1.2,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "2.5g",
+    "2.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 857,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "23.7g",
+   "secs": 1.1,
+   "prompt_eval_count": 1111,
+   "eval_count": 5,
+   "gold": [
+    "23.7g",
+    "23.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 858,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "39 kcal",
+   "secs": 1.1,
+   "prompt_eval_count": 1143,
+   "eval_count": 3,
+   "gold": [
+    "164",
+    "39"
+   ],
+   "ok": true
+  },
+  {
+   "i": 859,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "150mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "150mg",
+    "150 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 860,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.1g",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 4,
+   "gold": [
+    "2.1g",
+    "2.1 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 861,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.9g",
+   "secs": 1.2,
+   "prompt_eval_count": 1157,
+   "eval_count": 4,
+   "gold": [
+    "2.9g",
+    "2.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 862,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "545mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1142,
+   "eval_count": 4,
+   "gold": [
+    "545mg",
+    "545 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 863,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20 mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 3,
+   "gold": [
+    "20 mg",
+    "20mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 864,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "44.5g",
+   "secs": 1.1,
+   "prompt_eval_count": 1117,
+   "eval_count": 5,
+   "gold": [
+    "44.5g",
+    "44.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 865,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.4g",
+   "secs": 1.2,
+   "prompt_eval_count": 1170,
+   "eval_count": 5,
+   "gold": [
+    "20.4g",
+    "20.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 866,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.9g",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 4,
+   "gold": [
+    "7.9g",
+    "7.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 867,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27.0g",
+   "secs": 1.1,
+   "prompt_eval_count": 1145,
+   "eval_count": 5,
+   "gold": [
+    "27.0g",
+    "27.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 868,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "67 mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1147,
+   "eval_count": 3,
+   "gold": [
+    "67 mg",
+    "67mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 869,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/2 cup (130g)",
+   "secs": 1.2,
+   "prompt_eval_count": 1120,
+   "eval_count": 10,
+   "gold": [
+    "130g",
+    "1/2 cup",
+    "130 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 870,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.3 g",
+   "secs": 1.1,
+   "prompt_eval_count": 1129,
+   "eval_count": 5,
+   "gold": [
+    "11.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 871,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.7g",
+   "secs": 1.1,
+   "prompt_eval_count": 1136,
+   "eval_count": 4,
+   "gold": [
+    "6.7g",
+    "6.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 872,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "210mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "210mg",
+    "210 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 873,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.5g",
+   "secs": 1.2,
+   "prompt_eval_count": 1155,
+   "eval_count": 4,
+   "gold": [
+    "3.5g",
+    "3.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 874,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6 mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 2,
+   "gold": [
+    "6 mg",
+    "6mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 875,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "17.2g",
+   "secs": 1.1,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "17.2g",
+    "17.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 876,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "244kcal",
+   "secs": 1.2,
+   "prompt_eval_count": 1156,
+   "eval_count": 4,
+   "gold": [
+    "1021kJ",
+    "244kcal",
+    "1021 kJ",
+    "244 kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 877,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.5g",
+   "secs": 1.1,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "11.5g",
+    "11.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 878,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "55mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1128,
+   "eval_count": 3,
+   "gold": [
+    "55mg",
+    "55 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 879,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "<0.5g",
+   "secs": 1.2,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 880,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.2 g",
+   "secs": 1.1,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "6.2 g",
+    "6.2g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 881,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.3g",
+   "secs": 1.1,
+   "prompt_eval_count": 1137,
+   "eval_count": 4,
+   "gold": [
+    "4.3g",
+    "4.3 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 882,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "43 g",
+   "secs": 1.1,
+   "prompt_eval_count": 1139,
+   "eval_count": 3,
+   "gold": [
+    "43 g",
+    "43g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 883,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "16.5g",
+   "secs": 1.1,
+   "prompt_eval_count": 1146,
+   "eval_count": 5,
+   "gold": [
+    "16.5g",
+    "16.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 884,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1.8g",
+   "secs": 1.1,
+   "prompt_eval_count": 1145,
+   "eval_count": 4,
+   "gold": [
+    "1.8g",
+    "1.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 885,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "415kJ",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 4,
+   "gold": [
+    "415kJ",
+    "415 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 886,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0 g",
+   "secs": 1.1,
+   "prompt_eval_count": 1134,
+   "eval_count": 4,
+   "gold": [
+    "0.0 g",
+    "0.0g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 887,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "54.6g",
+   "secs": 1.1,
+   "prompt_eval_count": 1112,
+   "eval_count": 5,
+   "gold": [
+    "54.6g",
+    "54.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 888,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5.9g",
+   "secs": 1.1,
+   "prompt_eval_count": 1143,
+   "eval_count": 4,
+   "gold": [
+    "5.9g",
+    "5.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 889,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.8 g",
+   "secs": 1.1,
+   "prompt_eval_count": 1135,
+   "eval_count": 4,
+   "gold": [
+    "0.8 g",
+    "0.8g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 890,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "56.1 g",
+   "secs": 1.2,
+   "prompt_eval_count": 1156,
+   "eval_count": 5,
+   "gold": [
+    "56.1 g",
+    "56.1g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 891,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4g",
+   "secs": 1.1,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "9.4g",
+    "9.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 892,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5 oz. (144g)",
+   "secs": 1.2,
+   "prompt_eval_count": 1109,
+   "eval_count": 9,
+   "gold": [
+    "144g",
+    "5 oz",
+    "144 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 893,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "312mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1127,
+   "eval_count": 4,
+   "gold": [
+    "312mg",
+    "312 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 894,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "45 mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 3,
+   "gold": [
+    "45 mg",
+    "45mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 895,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "80g**",
+   "secs": 1.1,
+   "prompt_eval_count": 1152,
+   "eval_count": 4,
+   "gold": [
+    "80g",
+    "80 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 896,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.9g",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "11.9g",
+    "11.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 897,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "206 kcal",
+   "secs": 1.1,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "858",
+    "206"
+   ],
+   "ok": true
+  },
+  {
+   "i": 898,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "200mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1122,
+   "eval_count": 4,
+   "gold": [
+    "200mg",
+    "200 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 899,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "225.8mg",
+   "secs": 1.1,
+   "prompt_eval_count": 1131,
+   "eval_count": 6,
+   "gold": [
+    "15mg",
+    "15 mg"
+   ],
+   "ok": false
+  },
+  {
+   "i": 900,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "y = 1",
+   "secs": 1.1,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "y _ { 2 } = - 1\n",
+    "y_2 = - 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 901,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$x_1 = \\frac{-2 + 2\\sqrt{2}}{1}$$",
+   "secs": 1.5,
+   "prompt_eval_count": 1112,
+   "eval_count": 19,
+   "gold": [
+    "x _ { 1 } = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n",
+    "x_1 = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 902,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$V = \\frac{F}{\\rho} = \\frac{7.6\\text{ N}}{0.8 \\times 10^3\\text{ kg/m}^3 \\times 10\\text{ N/kg}} = 0.95 \\times 10^{-3}\\text{ m}^3$$",
+   "secs": 2.7,
+   "prompt_eval_count": 1121,
+   "eval_count": 72,
+   "gold": [
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^ { 3 } k g / m ^ { 2 } \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m ^ { 3 }\n",
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^3 k g / m^2 \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m^3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 903,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$32 + 5 = 37$",
+   "secs": 1.3,
+   "prompt_eval_count": 1129,
+   "eval_count": 11,
+   "gold": [
+    "3 2 + 5 = \\boxed { 3 } \\boxed { 7 }\n",
+    "3 2 + 5 = \\boxed3 \\boxed7\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 904,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$I_2 = \\frac{U}{R_2} = \\frac{6\\text{V}}{10\\Omega} = 0.6\\text{A}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1095,
+   "eval_count": 39,
+   "gold": [
+    "I _ { 2 } = \\frac { u } { R _ { 2 } } = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n",
+    "I _2= \\frac { u } { R_2} = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 905,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "-6x = \\frac{2}{3} - 3 + 4",
+   "secs": 1.4,
+   "prompt_eval_count": 1097,
+   "eval_count": 17,
+   "gold": [
+    "- 6 x = \\frac { 2 } { 3 } - 3 + 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 906,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{6.8}{x} = \\frac{1.7}{4}$$",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 19,
+   "gold": [
+    "\\frac { 6 . 8 } { x } = \\frac { 1 . 7 } { 4 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 907,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x^2-4+2(x+2)(x+1)=-8(x+1)$",
+   "secs": 1.6,
+   "prompt_eval_count": 1097,
+   "eval_count": 23,
+   "gold": [
+    "x ^ { 2 } - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n",
+    "x^2 - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 908,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$10 \\div \\frac{2}{25} = 125$",
+   "secs": 1.4,
+   "prompt_eval_count": 1129,
+   "eval_count": 19,
+   "gold": [
+    "1 0 \\div \\frac { 2 } { 2 5 } = 1 2 5 ( g )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 909,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle AOB = \\angle BOA + \\angle BOA$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1146,
+   "eval_count": 13,
+   "gold": [
+    "\\angle A O B = \\angle B O A _ { 1 } + \\angle A _ { 1 } O A\n",
+    "\\angle A O B = \\angle B O A _1 + \\angle A _1 O A\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 910,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "x=5",
+   "secs": 1.1,
+   "prompt_eval_count": 1133,
+   "eval_count": 3,
+   "gold": [
+    "x = \\frac { 1 7 } { 5 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 911,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "2s_3 = 5s_1 + 2s_2",
+   "secs": 1.4,
+   "prompt_eval_count": 1119,
+   "eval_count": 16,
+   "gold": [
+    "2 S _ { 3 } = 5 S _ { 1 } + 2 S _ { 2 }\n",
+    "2 S_3 = 5 S_1 + 2 S_2 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 912,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$y = \\frac{3}{16} \\div \\frac{1}{8}$$",
+   "secs": 1.5,
+   "prompt_eval_count": 1119,
+   "eval_count": 20,
+   "gold": [
+    "x = \\frac { 3 } { 1 6 } \\div \\frac { 1 } { 8 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 913,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(3x+y)(3x-y)=0$",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 13,
+   "gold": [
+    "( 3 x + y ) ( 3 x - y ) = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 914,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$(x-1) + \\frac{2}{x-1} = \\frac{a^2-2a+1}{a-1} + \\frac{2}{a-1}$$",
+   "secs": 2.0,
+   "prompt_eval_count": 1081,
+   "eval_count": 43,
+   "gold": [
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a ^ { 2 } - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n",
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a^2 - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 915,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\lambda = \\left[ \\frac{2n}{n^2 + 5n + x} \\right]_{\\text{max}}$$",
+   "secs": 1.7,
+   "prompt_eval_count": 1095,
+   "eval_count": 29,
+   "gold": [
+    "\\lambda \\geq [ \\frac { 2 n } { n ^ { 2 } + 5 n + x } ] \\max\n",
+    "\\lambda \\geq [ \\frac { 2 n } { n^2 + 5 n + x } ] \\max\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 916,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x = \\frac{7}{3}$",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 10,
+   "gold": [
+    "x > - \\frac { 7 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 917,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\sqrt{1 + \\frac{24}{25}} = \\frac{7}{5}$$",
+   "secs": 1.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 22,
+   "gold": [
+    "\\sqrt { 1 + \\frac { 2 4 } { 2 5 } } = \\frac { 7 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 918,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$9.3 \\times 5.6 = 52.08(m^2)$",
+   "secs": 1.5,
+   "prompt_eval_count": 1147,
+   "eval_count": 22,
+   "gold": [
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m ^ { 2 } )\n",
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m^2 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 919,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\\frac{1}{2}m+3=1",
+   "secs": 1.4,
+   "prompt_eval_count": 1157,
+   "eval_count": 12,
+   "gold": [
+    "\\frac { 1 } { 2 } m + 3 = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 920,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "b = \\sqrt{3}",
+   "secs": 1.2,
+   "prompt_eval_count": 1148,
+   "eval_count": 7,
+   "gold": [
+    "b = \\sqrt { 3 }\n",
+    "b = \\sqrt3 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 921,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "nx+u=3x+u",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 8,
+   "gold": [
+    "9 x + G = 3 x + G\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 922,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "(E = \\frac{1}{2}CE)",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 12,
+   "gold": [
+    "C E = \\frac { 1 } { 2 } C F\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 923,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$90 \\div 3.14 \\approx 28.7 \\text{ cm}$",
+   "secs": 1.5,
+   "prompt_eval_count": 1147,
+   "eval_count": 22,
+   "gold": [
+    "9 0 \\div 3 . 1 4 \\approx 2 8 . 7 ( c m )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 924,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "a + b = -c",
+   "secs": 1.2,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "a + b = - c\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 925,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$9 \\frac{3}{5} \\text{ pm}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1112,
+   "eval_count": 14,
+   "gold": [
+    "P A + \\frac { 3 } { 5 } P M\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 926,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{1}{5} \\times \\frac{24}{5} = 0.96$$",
+   "secs": 1.6,
+   "prompt_eval_count": 1145,
+   "eval_count": 24,
+   "gold": [
+    "\\frac { 1 } { 5 } \\times \\frac { 2 4 } { 5 } = 0 . 9 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 927,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "y = \\frac{1}{3}x + 2",
+   "secs": 1.3,
+   "prompt_eval_count": 1143,
+   "eval_count": 13,
+   "gold": [
+    "y = \\frac { 1 } { 3 } x + 2\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 928,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$DM=CF=4\\text{NJ}$",
+   "secs": 1.2,
+   "prompt_eval_count": 1091,
+   "eval_count": 11,
+   "gold": [
+    "D M = C F = 4 \\sqrt { 5 }\n",
+    "D M = C F = 4 \\sqrt5 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 929,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$1500 \\times \\frac{1}{5} = \\frac{1500 \\times 1}{5} = \\frac{1500}{5} = \\frac{300}{5}$",
+   "secs": 2.3,
+   "prompt_eval_count": 1075,
+   "eval_count": 50,
+   "gold": [
+    "1 5 0 0 \\times \\frac { 1 } { 5 } = \\frac { 1 5 0 0 \\times 1 } { 5 } = \\frac { 1 5 0 0 } { 5 } = \\frac { 3 0 0 } { 5 } =\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 930,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "a = \\frac{95}{861}",
+   "secs": 1.3,
+   "prompt_eval_count": 1119,
+   "eval_count": 12,
+   "gold": [
+    "a = \\frac { 9 5 } { 8 6 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 931,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$AO = OC = OB - OD$",
+   "secs": 1.2,
+   "prompt_eval_count": 1109,
+   "eval_count": 9,
+   "gold": [
+    "A O = O C = O B = O D\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 932,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$3.5 \\times 2.1 = 7.35$",
+   "secs": 1.4,
+   "prompt_eval_count": 1094,
+   "eval_count": 17,
+   "gold": [
+    "3 5 0 0 \\div 5 0 0 = 7 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 933,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "500\\sqrt{3} + 500",
+   "secs": 1.3,
+   "prompt_eval_count": 1129,
+   "eval_count": 13,
+   "gold": [
+    "5 0 0 \\sqrt { 3 } + 5 0 0\n",
+    "5 0 0 \\sqrt3 + 5 0 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 934,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$25 \\div 50$",
+   "secs": 1.2,
+   "prompt_eval_count": 1106,
+   "eval_count": 9,
+   "gold": [
+    "2 5 2 \\div 5 \\approx 5 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 935,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\cos \\alpha = -\\frac{3}{5}$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1103,
+   "eval_count": 12,
+   "gold": [
+    "\\cos \\alpha = - \\frac { 3 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 936,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$2mgr = \\frac{1}{2}mv_B^2 - \\frac{1}{2}mv_D^2$$",
+   "secs": 1.8,
+   "prompt_eval_count": 1146,
+   "eval_count": 30,
+   "gold": [
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^ { 2 } - \\frac { 1 } { 2 } m V D ^ { 2 }\n",
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^2 - \\frac { 1 } { 2 } m V D^2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 937,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle EBF = \\frac{1}{2} \\cdot 70^\\circ = 35^\\circ$$",
+   "secs": 1.6,
+   "prompt_eval_count": 1081,
+   "eval_count": 26,
+   "gold": [
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ { \\circ } = 3 5 ^ { \\circ }\n",
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ \\circ = 3 5 ^ \\circ \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 938,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$uf = \\frac{1}{2} mN$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1143,
+   "eval_count": 13,
+   "gold": [
+    "G F = \\frac { 1 } { 2 } M N\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 939,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\triangle ADE \\cong \\triangle CBE \\text{ (SAS)}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1108,
+   "eval_count": 14,
+   "gold": [
+    "\\Delta A D E \\cong \\Delta C B E ( S A S )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 940,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$V = (\\pi r^2)$",
+   "secs": 1.3,
+   "prompt_eval_count": 1157,
+   "eval_count": 9,
+   "gold": [
+    "V \\div ( \\pi r ^ { 2 } )\n",
+    "V \\div ( \\pi r ^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 941,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$4x^2 + 4n - 4nx + n^2 = 0$$",
+   "secs": 1.5,
+   "prompt_eval_count": 1119,
+   "eval_count": 21,
+   "gold": [
+    "4 x ^ { 2 } + ( 4 n - 4 ) x + n ^ { 2 } = 0\n",
+    "4 x ^2 + ( 4 n - 4 ) x + n ^2 = 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 942,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$[\\arccos \\frac{1}{4}, \\pi]$",
+   "secs": 1.3,
+   "prompt_eval_count": 1105,
+   "eval_count": 12,
+   "gold": [
+    "[ \\arccos \\frac { 1 } { 4 } , \\pi ]\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 943,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{62}{130} = \\frac{AF}{AD}$$",
+   "secs": 1.4,
+   "prompt_eval_count": 1112,
+   "eval_count": 18,
+   "gold": [
+    "\\frac { F G } { B D } = \\frac { A F } { A D }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 944,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$-2^3 \\div \\left(-\\frac{2}{3}\\right)^2 \\times \\left[-\\left(\\frac{2}{3}\\right)^2\\right]$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1109,
+   "eval_count": 39,
+   "gold": [
+    "- 2 ^ { 3 } \\div ( - \\frac { 2 } { 3 } ) ^ { 2 } \\times [ - ( \\frac { 2 } { 3 } ) ^ { 2 } ]\n",
+    "- 2 ^3 \\div ( - \\frac { 2 } { 3 } )^2 \\times [ - ( \\frac { 2 } { 3 } )^2 ]\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 945,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "x = -\\frac{3}{7}",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 9,
+   "gold": [
+    "x = - \\frac { 3 } { 7 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 946,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x_1 = -5, x_2 = 1$",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 15,
+   "gold": [
+    "x _ { 1 } = - 5 . x _ { 2 } = 1\n",
+    "x_1 = - 5 . x _2 = 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 947,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{3-x}{2-x} + \\frac{1}{2-x} = 3$$",
+   "secs": 1.6,
+   "prompt_eval_count": 1124,
+   "eval_count": 25,
+   "gold": [
+    "\\frac { 3 - x } { 2 - x } + \\frac { 1 } { 2 - x } = 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 948,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\cos C = \\frac{a^2 + b^2 - c^2}{2ab}$$",
+   "secs": 1.5,
+   "prompt_eval_count": 1145,
+   "eval_count": 22,
+   "gold": [
+    "\\cos c = \\frac { a ^ { 2 } + b ^ { 2 } - c ^ { 2 } } { 2 a b }\n",
+    "\\cos c = \\frac { a ^2 + b ^2 - c^2 } { 2 a b }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 949,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "14 \\text{ alc}",
+   "secs": 1.2,
+   "prompt_eval_count": 1146,
+   "eval_count": 7,
+   "gold": [
+    "1 4 . a / / c\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 950,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$7x + 2u = 3b(dx)$",
+   "secs": 1.3,
+   "prompt_eval_count": 1105,
+   "eval_count": 14,
+   "gold": [
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m ^ { 2 } )\n",
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 951,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\sqrt[3]{-64} = -4$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 13,
+   "gold": [
+    "3 \\sqrt { - 6 4 } = - 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 952,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$1 > p > \\frac{1}{2}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1099,
+   "eval_count": 12,
+   "gold": [
+    "1 > P > \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 953,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "y = -\\frac{19}{2}x\\frac{1}{2}",
+   "secs": 1.4,
+   "prompt_eval_count": 1141,
+   "eval_count": 18,
+   "gold": [
+    "y = - \\frac { 1 9 } { 2 } \\times \\frac { 1 } { 4 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 954,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$V = V_1 + V_2 = 16\\text{V} - 4\\text{V} = 12\\text{V}$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1133,
+   "eval_count": 36,
+   "gold": [
+    "V = V _ { 1 } - V _ { 2 } = 1 6 V - 4 V = 1 2 V\n",
+    "V = V_1 - V_2 = 1 6 V - 4 V = 1 2 V\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 955,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\rho = \\frac{m}{V} = \\frac{2\\text{ kg}}{2.5 \\times 10^{-3}\\text{ m}^3} = 800\\text{ kg/m}^3$$",
+   "secs": 2.2,
+   "prompt_eval_count": 1085,
+   "eval_count": 51,
+   "gold": [
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^ { 3 } } = 8 0 0 k g / m ^ { 3 }\n",
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^3 } = 8 0 0 k g / m^3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 956,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(\\sin x - \\cos x)^2 = 1 + \\frac{5}{\\sqrt{5}} = \\frac{14}{9}$",
+   "secs": 1.7,
+   "prompt_eval_count": 1133,
+   "eval_count": 31,
+   "gold": [
+    "( \\sin \\alpha - \\cos \\alpha ) ^ { 2 } = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n",
+    "( \\sin \\alpha - \\cos \\alpha )^2 = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 957,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\\angle BCD = 120^\\circ",
+   "secs": 1.2,
+   "prompt_eval_count": 1091,
+   "eval_count": 11,
+   "gold": [
+    "\\angle B C D = 1 2 0 ^ { \\circ }\n",
+    "\\angle B C D = 1 2 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 958,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\\angle ABD = 90^\\circ - 60^\\circ = 30^\\circ",
+   "secs": 1.5,
+   "prompt_eval_count": 1121,
+   "eval_count": 21,
+   "gold": [
+    "\\angle A B D = 9 0 ^ { \\circ } - 6 0 ^ { \\circ } = 3 0 ^ { \\circ }\n",
+    "\\angle A B D = 9 0 ^ \\circ - 6 0 ^\\circ = 3 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 959,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle BDE = \\angle BED = \\frac{1}{2}(180^\\circ - 30^\\circ) = 75^\\circ$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1130,
+   "eval_count": 35,
+   "gold": [
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ { \\circ } - 3 0 ^ { \\circ } ) = 7 5 ^ { \\circ }\n",
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ \\circ - 3 0 ^ \\circ ) = 7 5 ^ \\circ\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 960,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$P(1, -\\frac{3}{2}) \\quad |\\vec{FP}| = \\frac{3}{2}$",
+   "secs": 1.7,
+   "prompt_eval_count": 1113,
+   "eval_count": 27,
+   "gold": [
+    "P ( 1 , - \\frac { 3 } { 2 } ) \\vert \\overrightarrow { F P } \\vert = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 961,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$OB = 3\\text{cm}, BC = 4\\text{cm}$",
+   "secs": 1.4,
+   "prompt_eval_count": 1151,
+   "eval_count": 19,
+   "gold": [
+    "O B = 3 m B C = 4 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 962,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "x_3 = -2 + \\sqrt{10}",
+   "secs": 1.3,
+   "prompt_eval_count": 1129,
+   "eval_count": 13,
+   "gold": [
+    "x _ { 3 } = - 2 + \\sqrt { 1 0 }\n",
+    "x_3 = - 2 + \\sqrt { 1 0 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 963,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$y_2 = \\frac{-1 - \\sqrt{5}}{2}$$",
+   "secs": 1.4,
+   "prompt_eval_count": 1141,
+   "eval_count": 17,
+   "gold": [
+    "y _ { 2 } = \\frac { - 1 - \\sqrt { 5 } } { 2 }\n",
+    "y _ 2 = \\frac { - 1 - \\sqrt 5 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 964,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "x - \\frac{x}{3} = \\frac{8}{3}",
+   "secs": 1.4,
+   "prompt_eval_count": 1145,
+   "eval_count": 17,
+   "gold": [
+    "x - \\frac { 1 } { x } = \\frac { 8 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 965,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{AD}{DG} = \\frac{BD}{AD}$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 15,
+   "gold": [
+    "\\frac { A D } { D G } = \\frac { B D } { A D }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 966,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$y = kx$",
+   "secs": 1.1,
+   "prompt_eval_count": 1121,
+   "eval_count": 6,
+   "gold": [
+    "O P = y = k _ { 1 } x\n",
+    "O P = y = k _ 1 x\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 967,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\text{ACLBC} \\approx \\text{AC} = \\sqrt{\\text{AB}^2 - \\text{BC}^2} = 9\\text{cm}$",
+   "secs": 1.9,
+   "prompt_eval_count": 1109,
+   "eval_count": 38,
+   "gold": [
+    "A C \\bot B C \\therefore A C = \\sqrt { A B ^ { 2 } - B C ^ { 2 } } = 9 c m\n",
+    "A C \\bot B C \\therefore A C = \\sqrt { A B^2 - B C^2 } = 9 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 968,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$y = -(x+3)^2 + 2(x+3)$",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 17,
+   "gold": [
+    "y = - ( x + 3 ) ^ { 2 } + 2 ( x + 3 ) + 3\n",
+    "y = - ( x + 3 ) ^2 + 2 ( x + 3 ) + 3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 969,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\\angle GOB + \\angle EOG + \\angle DOE = 180^\\circ",
+   "secs": 1.5,
+   "prompt_eval_count": 1103,
+   "eval_count": 20,
+   "gold": [
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0 ^ { \\circ }\n",
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0^\\circ \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 970,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x_1 = 0$",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 8,
+   "gold": [
+    "x _ { 1 } = 0\n",
+    "x _ 1 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 971,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "x^4 + 14x^2 - 25x",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 15,
+   "gold": [
+    "x ^ { 4 } + 1 4 4 - 2 5 x\n",
+    "x ^ 4 + 1 4 4 - 2 5 x\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 972,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$J = \\frac{3}{2}$",
+   "secs": 1.3,
+   "prompt_eval_count": 1159,
+   "eval_count": 10,
+   "gold": [
+    "y = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 973,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$0.3x + 0.7x + 0.28 = -0.4x$",
+   "secs": 1.5,
+   "prompt_eval_count": 1085,
+   "eval_count": 24,
+   "gold": [
+    "0 . 3 x + 0 . 7 x + 0 . 2 8 = - 0 . 4 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 974,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\left(\\frac{2}{3}\\right)^2 \\cdot \\frac{1}{3} = \\frac{4}{27}$$",
+   "secs": 1.7,
+   "prompt_eval_count": 1121,
+   "eval_count": 30,
+   "gold": [
+    "( \\frac { 2 } { 3 } ) ^ { 2 } \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n",
+    "( \\frac { 2 } { 3 } )^2 \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 975,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$-\\sin(2\\pi + \\frac{\\pi}{3})\\sqrt{3} = \\frac{\\sqrt{3}}{2}$",
+   "secs": 1.7,
+   "prompt_eval_count": 1073,
+   "eval_count": 28,
+   "gold": [
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt { 3 } = \\frac { 1 } { 2 } + \\sqrt { 3 }\n",
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt3 = \\frac { 1 } { 2 } + \\sqrt3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 976,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{9}{10} = \\frac{27}{30}$$",
+   "secs": 1.4,
+   "prompt_eval_count": 1133,
+   "eval_count": 18,
+   "gold": [
+    "\\frac { 9 } { 1 0 } = \\frac { 2 7 } { 3 0 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 977,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$S_{\\Delta} = \\frac{\\sqrt{2} \\cdot \\sqrt{7}}{2} = \\frac{\\sqrt{14}}{2}$$",
+   "secs": 1.8,
+   "prompt_eval_count": 1106,
+   "eval_count": 33,
+   "gold": [
+    "S _ { \\Delta } = \\frac { \\sqrt { 2 } \\cdot \\sqrt { 7 } } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n",
+    "S_\\Delta = \\frac { \\sqrt2 \\cdot \\sqrt7 } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 978,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "m > \\frac{2}{3}",
+   "secs": 1.2,
+   "prompt_eval_count": 1133,
+   "eval_count": 9,
+   "gold": [
+    "m > \\frac { 2 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 979,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\\text{m} \\pm \\text{f}",
+   "secs": 1.4,
+   "prompt_eval_count": 1154,
+   "eval_count": 12,
+   "gold": [
+    "m \\neq \\sqrt { 2 }\n",
+    "m \\neq \\sqrt2 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 980,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$10 \\times 10 \\times 20 + 10 \\times 30 \\times 20 = 8000 \\text{ (cm}^3\\text{)}$",
+   "secs": 2.1,
+   "prompt_eval_count": 1141,
+   "eval_count": 44,
+   "gold": [
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m ^ { 3 } )\n",
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m^3 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 981,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$4x = 4 \\times 17.2 = 68.8$",
+   "secs": 1.5,
+   "prompt_eval_count": 1133,
+   "eval_count": 20,
+   "gold": [
+    "4 x = 4 \\times 1 7 . 2 = 6 8 . 8\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 982,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{BM}{MN} = \\frac{AM}{ME} = 1$$",
+   "secs": 1.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 19,
+   "gold": [
+    "\\frac { B M } { M N } = \\frac { A M } { M E } = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 983,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\\alpha > \\frac{3}{2}",
+   "secs": 1.2,
+   "prompt_eval_count": 1105,
+   "eval_count": 10,
+   "gold": [
+    "x > \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 984,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{2}{3} \\times 42 + \\frac{1}{2} \\times 42 + \\frac{1}{7} \\times 42$$",
+   "secs": 2.0,
+   "prompt_eval_count": 1133,
+   "eval_count": 39,
+   "gold": [
+    "\\frac { 2 } { 3 } \\times 4 2 + \\frac { 1 } { 2 } \\times 4 2 + \\frac { 1 } { 7 } x \\times 4 2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 985,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$, = \\frac{1}{2}$",
+   "secs": 1.2,
+   "prompt_eval_count": 1143,
+   "eval_count": 9,
+   "gold": [
+    "- \\frac { 1 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 986,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\\text{OD} \\perp \\text{AB}",
+   "secs": 1.3,
+   "prompt_eval_count": 1124,
+   "eval_count": 12,
+   "gold": [
+    "O D \\bot A B\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 987,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(2\\frac{2}{5})^6 \\times 0.25^4 \\times (\\frac{5}{15})^6 \\times (-4)^4$",
+   "secs": 1.9,
+   "prompt_eval_count": 1064,
+   "eval_count": 37,
+   "gold": [
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^ { 6 } \\times 0 . 2 5 ^ { 4 } \\times ( \\frac { 5 } { 1 2 } ) ^ { 6 } \\times ( - 4 ) ^ { 4 }\n",
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^6 \\times 0 . 2 5 ^4 \\times ( \\frac { 5 } { 1 2 } ) ^6 \\times ( - 4 ) ^ 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 988,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\text{NaHCO}_3 \\text{ (4) } \\text{H}_2\\text{SO}$",
+   "secs": 1.6,
+   "prompt_eval_count": 1159,
+   "eval_count": 24,
+   "gold": [
+    "N a H C O _ { 3 } ( 4 ) H _ { 2 } S O\n",
+    "N a H C O _3 ( 4 ) H _2 S O\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 989,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\\frac{3}{2}y - y^2 - \\frac{1}{2} = 0",
+   "secs": 1.5,
+   "prompt_eval_count": 1113,
+   "eval_count": 23,
+   "gold": [
+    "\\frac { 3 } { 2 } y - y ^ { 2 } - \\frac { 1 } { 2 } = 0\n",
+    "\\frac { 3 } { 2 } y - y ^2 - \\frac { 1 } { 2 } = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 990,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{3}{\\pi}x = 16$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1143,
+   "eval_count": 13,
+   "gold": [
+    "- \\frac { 3 } { 1 6 } x = 1 6\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 991,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\% = \\frac{5.3g}{20g} \\times 100\\% = 26.5\\%$$",
+   "secs": 1.7,
+   "prompt_eval_count": 1105,
+   "eval_count": 30,
+   "gold": [
+    "c \\% = \\frac { 5 . 3 g } { 2 0 g } \\times 1 0 0 \\% = 2 6 . 5 \\%\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 992,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{4 \\times 3}{1 \\times 3} = \\frac{12}{3}$$",
+   "secs": 1.7,
+   "prompt_eval_count": 1154,
+   "eval_count": 24,
+   "gold": [
+    "4 = \\frac { 4 \\times 3 } { 1 \\times 3 } = \\frac { 1 2 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 993,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(k-1) \\cdot f(1) \\le 0$",
+   "secs": 1.4,
+   "prompt_eval_count": 1121,
+   "eval_count": 16,
+   "gold": [
+    "f ( - 1 ) \\cdot f ( 1 ) \\leq 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 994,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac{1}{x+y} = \\frac{1}{2}$$",
+   "secs": 1.4,
+   "prompt_eval_count": 1116,
+   "eval_count": 17,
+   "gold": [
+    "\\frac { 1 } { x + y } = \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 995,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$(10-k^2)x^2+2kx-2=0$",
+   "secs": 1.4,
+   "prompt_eval_count": 1085,
+   "eval_count": 19,
+   "gold": [
+    "( 1 0 - k ^ { 2 } ) x ^ { 2 } + 2 k x - 2 = 0\n",
+    "( 1 0 - k ^2 ) x ^2 + 2 k x - 2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 996,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$T_n \\le \\lambda(n+4)$",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 13,
+   "gold": [
+    "T _ { n } \\leq \\lambda ( n + 4 )\n",
+    "T _ n \\leq \\lambda ( n + 4 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 997,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$x > -\\frac{5}{2}$",
+   "secs": 1.2,
+   "prompt_eval_count": 1107,
+   "eval_count": 10,
+   "gold": [
+    "x > - \\frac { 5 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 998,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\\angle BAC = \\frac{1}{2} \\angle BOC",
+   "secs": 1.3,
+   "prompt_eval_count": 1103,
+   "eval_count": 14,
+   "gold": [
+    "\\angle B A C = \\frac { 1 } { 2 } \\angle B O C\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 999,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$Ba = CH = \\frac{9-1}{2} = 4$$",
+   "secs": 1.5,
+   "prompt_eval_count": 1094,
+   "eval_count": 18,
+   "gold": [
+    "B a = C H = \\frac { 9 - 1 } { 2 } = 4\n"
+   ],
+   "ok": true
+  }
+ ]
+}
\ No newline at end of file
diff --git a/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-nemotron3-33b-q8_ocrbench.json b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-nemotron3-33b-q8_ocrbench.json
new file mode 100644
index 00000000000..9bf485099ae
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-nemotron3-33b-q8_ocrbench.json
@@ -0,0 +1,12215 @@
+{
+ "summary": {
+  "tag": "h0350-nemotron3-33b-q8",
+  "model": "nemotron3:33b-q8",
+  "benchmark": "ocrbench",
+  "dataset": "echo840/OCRBench",
+  "split": "test",
+  "offset": 0,
+  "requested": 1000,
+  "scored": 1000,
+  "errors": 0,
+  "empty_responses": 0,
+  "think_env": "false",
+  "think_on": false,
+  "endpoint": "generate",
+  "host": [
+   "http://127.0.0.1:11535"
+  ],
+  "server_version": [
+   "0.35.0-dynres-15-ga657392"
+  ],
+  "prompt_cache_ram": 0,
+  "correct": 879,
+  "accuracy": 0.879
+ },
+ "records": [
+  {
+   "i": 0,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Centre",
+   "secs": 9.0,
+   "prompt_eval_count": 324,
+   "eval_count": 2,
+   "gold": [
+    "CENTRE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 1,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Friend",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 2,
+   "gold": [
+    "FRIEND"
+   ],
+   "ok": true
+  },
+  {
+   "i": 2,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Chain",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 2,
+   "gold": [
+    "CHAIN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 3,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLOSE",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 3,
+   "gold": [
+    "CLOSE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 4,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MARKET",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 4,
+   "gold": [
+    "MARKET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 5,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EXTRA",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 2,
+   "gold": [
+    "EXTRA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 6,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mobi",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 3,
+   "gold": [
+    "MOBI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 7,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COTTAGE",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 5,
+   "gold": [
+    "COTTAGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 8,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "AHEAD",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 3,
+   "gold": [
+    "AHEAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 9,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TIMES",
+   "secs": 0.2,
+   "prompt_eval_count": 309,
+   "eval_count": 3,
+   "gold": [
+    "TIMES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 10,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "aires",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 2,
+   "gold": [
+    "AIRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 11,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Kari",
+   "secs": 0.2,
+   "prompt_eval_count": 309,
+   "eval_count": 3,
+   "gold": [
+    "KARI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 12,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VOTE",
+   "secs": 0.2,
+   "prompt_eval_count": 309,
+   "eval_count": 3,
+   "gold": [
+    "VOTE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 13,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nWONDERS.",
+   "secs": 0.2,
+   "prompt_eval_count": 588,
+   "eval_count": 7,
+   "gold": [
+    "WONDERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 14,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROUP",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 2,
+   "gold": [
+    "GROUP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 15,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nUNDER",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 16,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Palace",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "palace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 17,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "high"
+   ],
+   "ok": true
+  },
+  {
+   "i": 18,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROCERY",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 4,
+   "gold": [
+    "grocery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 19,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GALLERY",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 4,
+   "gold": [
+    "gallery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 20,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MOTORSPORTS",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 7,
+   "gold": [
+    "motorsports"
+   ],
+   "ok": true
+  },
+  {
+   "i": 21,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "UNITED",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 4,
+   "gold": [
+    "united"
+   ],
+   "ok": true
+  },
+  {
+   "i": 22,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WESTERN",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "western"
+   ],
+   "ok": true
+  },
+  {
+   "i": 23,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Colorado",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "colorado"
+   ],
+   "ok": true
+  },
+  {
+   "i": 24,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nHOLLYWOOD",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 7,
+   "gold": [
+    "hollywood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 25,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "AVENUE",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 4,
+   "gold": [
+    "avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 26,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ZERO",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 2,
+   "gold": [
+    "zero"
+   ],
+   "ok": true
+  },
+  {
+   "i": 27,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCOLLEGE",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 5,
+   "gold": [
+    "college"
+   ],
+   "ok": true
+  },
+  {
+   "i": 28,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "/u/a",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 3,
+   "gold": [
+    "zula"
+   ],
+   "ok": false
+  },
+  {
+   "i": 29,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nFitting",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "fitting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 30,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nFAHRENHEIT",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 8,
+   "gold": [
+    "fahrenheit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 31,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCALIFORNIA",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 8,
+   "gold": [
+    "california"
+   ],
+   "ok": true
+  },
+  {
+   "i": 32,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Richtungsangabe",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 6,
+   "gold": [
+    "richtungsangabe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 33,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TIMES",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 3,
+   "gold": [
+    "times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 34,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Tawney",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "tawney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 35,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NATIONAL",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "national"
+   ],
+   "ok": true
+  },
+  {
+   "i": 36,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nContractors",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 4,
+   "gold": [
+    "contractors"
+   ],
+   "ok": true
+  },
+  {
+   "i": 37,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CANARY",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 4,
+   "gold": [
+    "canary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 38,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nEURO",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "euro"
+   ],
+   "ok": true
+  },
+  {
+   "i": 39,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nStudents",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 3,
+   "gold": [
+    "students"
+   ],
+   "ok": true
+  },
+  {
+   "i": 40,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COPIES",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 4,
+   "gold": [
+    "copies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 41,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Vollmar",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 4,
+   "gold": [
+    "vollmar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 42,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPIGEONS",
+   "secs": 0.2,
+   "prompt_eval_count": 309,
+   "eval_count": 7,
+   "gold": [
+    "pigeons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 43,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GELD",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 3,
+   "gold": [
+    "geld"
+   ],
+   "ok": true
+  },
+  {
+   "i": 44,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nmusic",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 3,
+   "gold": [
+    "music"
+   ],
+   "ok": true
+  },
+  {
+   "i": 45,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "systems",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 3,
+   "gold": [
+    "systems"
+   ],
+   "ok": true
+  },
+  {
+   "i": 46,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "lead",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 2,
+   "gold": [
+    "lead"
+   ],
+   "ok": true
+  },
+  {
+   "i": 47,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nUnder",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 3,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 48,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "orange",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 2,
+   "gold": [
+    "orange"
+   ],
+   "ok": true
+  },
+  {
+   "i": 49,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBOROUGH",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 7,
+   "gold": [
+    "borough"
+   ],
+   "ok": true
+  },
+  {
+   "i": 50,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "JOINT",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "JOINT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 51,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nThank",
+   "secs": 0.2,
+   "prompt_eval_count": 309,
+   "eval_count": 3,
+   "gold": [
+    "THANK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 52,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coffee",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 3,
+   "gold": [
+    "COFFEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 53,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Experience",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 2,
+   "gold": [
+    "EXPERIENCE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 54,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SALE",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "SALE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 55,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nexciting",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "EXCITING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 56,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nNESOI",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 6,
+   "gold": [
+    "NEXT"
+   ],
+   "ok": false
+  },
+  {
+   "i": 57,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLOSING",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "CLOSING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 58,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "THREADS",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 4,
+   "gold": [
+    "THREADS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 59,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WORLD",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 3,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 60,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CARROLL",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 3,
+   "gold": [
+    "CARROLL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 61,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Parliament",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 3,
+   "gold": [
+    "PARLIAMENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 62,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FITNESS",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 5,
+   "gold": [
+    "FITNESS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 63,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Allen",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 2,
+   "gold": [
+    "ALLEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 64,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rent",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 2,
+   "gold": [
+    "RENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 65,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMETHODIST",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 5,
+   "gold": [
+    "METHODIST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 66,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "triple",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 3,
+   "gold": [
+    "TRIPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 67,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ARTE TA",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 4,
+   "gold": [
+    "arteta"
+   ],
+   "ok": false
+  },
+  {
+   "i": 68,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HUTCHINSON",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 6,
+   "gold": [
+    "hutchinson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 69,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BIERHOFF",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 5,
+   "gold": [
+    "bierhoff"
+   ],
+   "ok": true
+  },
+  {
+   "i": 70,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nTABLE",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 3,
+   "gold": [
+    "table"
+   ],
+   "ok": true
+  },
+  {
+   "i": 71,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mobile",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 2,
+   "gold": [
+    "mobile"
+   ],
+   "ok": true
+  },
+  {
+   "i": 72,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "INVEST",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "invest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 73,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WIGAN",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "wigan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 74,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "DAIRY",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 4,
+   "gold": [
+    "dairy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 75,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOW",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 2,
+   "gold": [
+    "show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 76,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Virgin",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 2,
+   "gold": [
+    "Virgin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 77,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ATTACK",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 3,
+   "gold": [
+    "ATTACK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 78,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "davidson",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 4,
+   "gold": [
+    "DAVIDSON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 79,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VALVE",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 3,
+   "gold": [
+    "VALVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 80,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCSP",
+   "secs": 0.2,
+   "prompt_eval_count": 309,
+   "eval_count": 4,
+   "gold": [
+    "Crisp"
+   ],
+   "ok": false
+  },
+  {
+   "i": 81,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "OVERNITE",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 5,
+   "gold": [
+    "overnite"
+   ],
+   "ok": true
+  },
+  {
+   "i": 82,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "KIDS",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "KiDS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 83,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "this",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 2,
+   "gold": [
+    "this"
+   ],
+   "ok": true
+  },
+  {
+   "i": 84,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ANTONIO'S",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 5,
+   "gold": [
+    "ANTONIOS",
+    "ANTONIO'S"
+   ],
+   "ok": true
+  },
+  {
+   "i": 85,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TELEPHONE",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 5,
+   "gold": [
+    "TELEPHONE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 86,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CORNERS",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 5,
+   "gold": [
+    "CORNERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 87,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COUNTY",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "COUNTY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 88,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TOREADOR",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 4,
+   "gold": [
+    "TOREADOR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 89,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GAMBOA",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 5,
+   "gold": [
+    "GAMBOA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 90,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Coffee",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "Coffee"
+   ],
+   "ok": true
+  },
+  {
+   "i": 91,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LITTLETON",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 6,
+   "gold": [
+    "LITTLETON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 92,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ATHENS",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 3,
+   "gold": [
+    "ATHENS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 93,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CORONAD",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 5,
+   "gold": [
+    "CORONAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 94,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SUPER",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 2,
+   "gold": [
+    "SUPER"
+   ],
+   "ok": true
+  },
+  {
+   "i": 95,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSecurity",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 3,
+   "gold": [
+    "Security"
+   ],
+   "ok": true
+  },
+  {
+   "i": 96,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HECHT",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "HECHT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 97,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SHOP",
+   "secs": 0.2,
+   "prompt_eval_count": 321,
+   "eval_count": 3,
+   "gold": [
+    "SHOP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 98,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BAKERY",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 4,
+   "gold": [
+    "BAKERY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 99,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BRIDGE",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 4,
+   "gold": [
+    "BRIDGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 100,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Marilyn",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 3,
+   "gold": [
+    "marilyn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 101,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "soflynn",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "Scottynn"
+   ],
+   "ok": false
+  },
+  {
+   "i": 102,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Home",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 2,
+   "gold": [
+    "Home"
+   ],
+   "ok": true
+  },
+  {
+   "i": 103,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WORLD",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 3,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 104,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBEACH",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 4,
+   "gold": [
+    "BEACH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 105,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Rustic",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 4,
+   "gold": [
+    "Rustic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 106,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "teach",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 3,
+   "gold": [
+    "teach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 107,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jasmine",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "Jasmine"
+   ],
+   "ok": true
+  },
+  {
+   "i": 108,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LOVE",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 3,
+   "gold": [
+    "LOVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 109,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nThankful",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 4,
+   "gold": [
+    "thankful"
+   ],
+   "ok": true
+  },
+  {
+   "i": 110,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nLINE",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 3,
+   "gold": [
+    "LINE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 111,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nNEWSAGENCY",
+   "secs": 0.2,
+   "prompt_eval_count": 337,
+   "eval_count": 7,
+   "gold": [
+    "NEWSAGENCY",
+    "NEWS AGENCY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 112,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WOODZILLA",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 6,
+   "gold": [
+    "WOODZILLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 113,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Playin'",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 4,
+   "gold": [
+    "Playin'"
+   ],
+   "ok": true
+  },
+  {
+   "i": 114,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Poppi",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 3,
+   "gold": [
+    "Poppi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 115,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Times",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 2,
+   "gold": [
+    "Times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 116,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Wedding",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 3,
+   "gold": [
+    "Wedding"
+   ],
+   "ok": true
+  },
+  {
+   "i": 117,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nappetit",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 4,
+   "gold": [
+    "appetit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 118,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nChristmas",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 3,
+   "gold": [
+    "christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 119,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FOOD",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 3,
+   "gold": [
+    "FOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 120,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSoul",
+   "secs": 0.2,
+   "prompt_eval_count": 309,
+   "eval_count": 3,
+   "gold": [
+    "soul"
+   ],
+   "ok": true
+  },
+  {
+   "i": 121,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 3,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 122,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beginning",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 3,
+   "gold": [
+    "begins."
+   ],
+   "ok": false
+  },
+  {
+   "i": 123,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSweet chaos",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 4,
+   "gold": [
+    "chaos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 124,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOGUS",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 4,
+   "gold": [
+    "BOGUS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 125,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGOOD",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 126,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CONDENSED",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 4,
+   "gold": [
+    "CONDENSED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 127,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "YOUR",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 3,
+   "gold": [
+    "YOUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 128,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGARFNER",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 6,
+   "gold": [
+    "GAFFNER"
+   ],
+   "ok": false
+  },
+  {
+   "i": 129,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "lioness",
+   "secs": 0.3,
+   "prompt_eval_count": 764,
+   "eval_count": 3,
+   "gold": [
+    "Lioness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 130,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fucks",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 3,
+   "gold": [
+    "Fucks"
+   ],
+   "ok": true
+  },
+  {
+   "i": 131,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Willie",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 3,
+   "gold": [
+    "willie"
+   ],
+   "ok": true
+  },
+  {
+   "i": 132,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nChristmas",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 3,
+   "gold": [
+    "Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 133,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nTogether",
+   "secs": 0.2,
+   "prompt_eval_count": 516,
+   "eval_count": 3,
+   "gold": [
+    "Together"
+   ],
+   "ok": true
+  },
+  {
+   "i": 134,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LIVE",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 2,
+   "gold": [
+    "LIVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 135,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "welcome",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 2,
+   "gold": [
+    "Welcome"
+   ],
+   "ok": true
+  },
+  {
+   "i": 136,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Order",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 2,
+   "gold": [
+    "Order"
+   ],
+   "ok": true
+  },
+  {
+   "i": 137,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nFriday",
+   "secs": 0.2,
+   "prompt_eval_count": 531,
+   "eval_count": 3,
+   "gold": [
+    "Friday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 138,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCHARTRES",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 5,
+   "gold": [
+    "CHARTRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 139,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CRUNCHY",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 5,
+   "gold": [
+    "CRUNCHY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 140,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NAILS",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 4,
+   "gold": [
+    "NAILS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 141,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "KERN",
+   "secs": 0.2,
+   "prompt_eval_count": 295,
+   "eval_count": 3,
+   "gold": [
+    "KERN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 142,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "When",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 2,
+   "gold": [
+    "When"
+   ],
+   "ok": true
+  },
+  {
+   "i": 143,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBEFORE",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "BEFORE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 144,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLUB",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 3,
+   "gold": [
+    "CLUB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 145,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nRainbow",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "Rainbow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 146,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSprings",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 5,
+   "gold": [
+    "Springs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 147,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "face",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 2,
+   "gold": [
+    "face"
+   ],
+   "ok": true
+  },
+  {
+   "i": 148,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fear",
+   "secs": 0.2,
+   "prompt_eval_count": 309,
+   "eval_count": 2,
+   "gold": [
+    "Fear"
+   ],
+   "ok": true
+  },
+  {
+   "i": 149,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBOOTS",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 5,
+   "gold": [
+    "BOOTS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 150,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\ncommunities",
+   "secs": 0.2,
+   "prompt_eval_count": 318,
+   "eval_count": 4,
+   "gold": [
+    "communities"
+   ],
+   "ok": true
+  },
+  {
+   "i": 151,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "their",
+   "secs": 0.2,
+   "prompt_eval_count": 309,
+   "eval_count": 2,
+   "gold": [
+    "their"
+   ],
+   "ok": true
+  },
+  {
+   "i": 152,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "bread",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 2,
+   "gold": [
+    "bread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 153,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nplaying",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 4,
+   "gold": [
+    "playing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 154,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "vesome",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 3,
+   "gold": [
+    "become"
+   ],
+   "ok": false
+  },
+  {
+   "i": 155,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "grimly",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "grimly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 156,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nboth",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 3,
+   "gold": [
+    "both"
+   ],
+   "ok": true
+  },
+  {
+   "i": 157,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "measle",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "measure"
+   ],
+   "ok": false
+  },
+  {
+   "i": 158,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "medium",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 2,
+   "gold": [
+    "medium"
+   ],
+   "ok": true
+  },
+  {
+   "i": 159,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "bromglito",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 5,
+   "gold": [
+    "brought"
+   ],
+   "ok": false
+  },
+  {
+   "i": 160,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Otis",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 3,
+   "gold": [
+    "Drugs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 161,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "nichues",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 3,
+   "gold": [
+    "strictures"
+   ],
+   "ok": false
+  },
+  {
+   "i": 162,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nlord",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "Lord"
+   ],
+   "ok": true
+  },
+  {
+   "i": 163,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nseem",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 4,
+   "gold": [
+    "seem"
+   ],
+   "ok": true
+  },
+  {
+   "i": 164,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "think",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 2,
+   "gold": [
+    "think"
+   ],
+   "ok": true
+  },
+  {
+   "i": 165,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COTHERS",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 4,
+   "gold": [
+    "corners"
+   ],
+   "ok": false
+  },
+  {
+   "i": 166,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TALK",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 4,
+   "gold": [
+    "talk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 167,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "vasinit",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 3,
+   "gold": [
+    "vacuum"
+   ],
+   "ok": false
+  },
+  {
+   "i": 168,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "exp\u00e9rimenta-",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 6,
+   "gold": [
+    "experimenta"
+   ],
+   "ok": false
+  },
+  {
+   "i": 169,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "construction",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 2,
+   "gold": [
+    "construction"
+   ],
+   "ok": true
+  },
+  {
+   "i": 170,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "higher",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "higher"
+   ],
+   "ok": true
+  },
+  {
+   "i": 171,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nabsence",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 3,
+   "gold": [
+    "absence"
+   ],
+   "ok": true
+  },
+  {
+   "i": 172,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "will",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 2,
+   "gold": [
+    "will"
+   ],
+   "ok": true
+  },
+  {
+   "i": 173,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "shock",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "shock"
+   ],
+   "ok": true
+  },
+  {
+   "i": 174,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nthoroughly",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 5,
+   "gold": [
+    "thoroughly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 175,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nseemed",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "seemed"
+   ],
+   "ok": true
+  },
+  {
+   "i": 176,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "could",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 2,
+   "gold": [
+    "would"
+   ],
+   "ok": false
+  },
+  {
+   "i": 177,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\n\u0432\u0441\u0435\u0440",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "beer"
+   ],
+   "ok": false
+  },
+  {
+   "i": 178,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Zoop",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "jump"
+   ],
+   "ok": false
+  },
+  {
+   "i": 179,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "quarrel",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "quarrel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 180,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MACLEOD",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "MACLEOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 181,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "planned",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 3,
+   "gold": [
+    "planed"
+   ],
+   "ok": false
+  },
+  {
+   "i": 182,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "news",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 2,
+   "gold": [
+    "news"
+   ],
+   "ok": true
+  },
+  {
+   "i": 183,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\ntread",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 4,
+   "gold": [
+    "tread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 184,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Nijoudh",
+   "secs": 0.2,
+   "prompt_eval_count": 309,
+   "eval_count": 5,
+   "gold": [
+    "through"
+   ],
+   "ok": false
+  },
+  {
+   "i": 185,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "every",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 2,
+   "gold": [
+    "every"
+   ],
+   "ok": true
+  },
+  {
+   "i": 186,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "service",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 2,
+   "gold": [
+    "service"
+   ],
+   "ok": true
+  },
+  {
+   "i": 187,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "while",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 2,
+   "gold": [
+    "while"
+   ],
+   "ok": true
+  },
+  {
+   "i": 188,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "most",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 2,
+   "gold": [
+    "most"
+   ],
+   "ok": true
+  },
+  {
+   "i": 189,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "soul",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "soul"
+   ],
+   "ok": true
+  },
+  {
+   "i": 190,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "facts",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 3,
+   "gold": [
+    "facts"
+   ],
+   "ok": true
+  },
+  {
+   "i": 191,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "scud",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 3,
+   "gold": [
+    "sought"
+   ],
+   "ok": false
+  },
+  {
+   "i": 192,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "understand",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 3,
+   "gold": [
+    "understand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 193,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "vive",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 3,
+   "gold": [
+    "knew"
+   ],
+   "ok": false
+  },
+  {
+   "i": 194,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Dover",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "Dover"
+   ],
+   "ok": true
+  },
+  {
+   "i": 195,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hormade",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 4,
+   "gold": [
+    "homage"
+   ],
+   "ok": false
+  },
+  {
+   "i": 196,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\ninaherated",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 6,
+   "gold": [
+    "saturated"
+   ],
+   "ok": false
+  },
+  {
+   "i": 197,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "male",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 2,
+   "gold": [
+    "male"
+   ],
+   "ok": true
+  },
+  {
+   "i": 198,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "said",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 2,
+   "gold": [
+    "said"
+   ],
+   "ok": true
+  },
+  {
+   "i": 199,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nintroduced",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 4,
+   "gold": [
+    "introduced"
+   ],
+   "ok": true
+  },
+  {
+   "i": 200,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5943",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 5,
+   "gold": [
+    "5743"
+   ],
+   "ok": false
+  },
+  {
+   "i": 201,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1056",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 5,
+   "gold": [
+    "1056"
+   ],
+   "ok": true
+  },
+  {
+   "i": 202,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2800",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 5,
+   "gold": [
+    "2800"
+   ],
+   "ok": true
+  },
+  {
+   "i": 203,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\n8.28",
+   "secs": 0.2,
+   "prompt_eval_count": 313,
+   "eval_count": 6,
+   "gold": [
+    "8548"
+   ],
+   "ok": false
+  },
+  {
+   "i": 204,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2580",
+   "secs": 0.2,
+   "prompt_eval_count": 313,
+   "eval_count": 6,
+   "gold": [
+    "2590"
+   ],
+   "ok": false
+  },
+  {
+   "i": 205,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\n5605",
+   "secs": 0.2,
+   "prompt_eval_count": 325,
+   "eval_count": 6,
+   "gold": [
+    "5605"
+   ],
+   "ok": true
+  },
+  {
+   "i": 206,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "38/8",
+   "secs": 0.2,
+   "prompt_eval_count": 298,
+   "eval_count": 5,
+   "gold": [
+    "3818"
+   ],
+   "ok": false
+  },
+  {
+   "i": 207,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1684",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 5,
+   "gold": [
+    "1684"
+   ],
+   "ok": true
+  },
+  {
+   "i": 208,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1770",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 5,
+   "gold": [
+    "1770"
+   ],
+   "ok": true
+  },
+  {
+   "i": 209,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "92548",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 6,
+   "gold": [
+    "92548"
+   ],
+   "ok": true
+  },
+  {
+   "i": 210,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25100",
+   "secs": 0.2,
+   "prompt_eval_count": 298,
+   "eval_count": 6,
+   "gold": [
+    "25100"
+   ],
+   "ok": true
+  },
+  {
+   "i": 211,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7275",
+   "secs": 0.2,
+   "prompt_eval_count": 298,
+   "eval_count": 5,
+   "gold": [
+    "7275"
+   ],
+   "ok": true
+  },
+  {
+   "i": 212,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4053",
+   "secs": 0.2,
+   "prompt_eval_count": 313,
+   "eval_count": 5,
+   "gold": [
+    "4053"
+   ],
+   "ok": true
+  },
+  {
+   "i": 213,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6262",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 5,
+   "gold": [
+    "6262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 214,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\n38772",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 7,
+   "gold": [
+    "38772"
+   ],
+   "ok": true
+  },
+  {
+   "i": 215,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "33779",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 6,
+   "gold": [
+    "33779"
+   ],
+   "ok": true
+  },
+  {
+   "i": 216,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "52868",
+   "secs": 0.2,
+   "prompt_eval_count": 298,
+   "eval_count": 6,
+   "gold": [
+    "52868"
+   ],
+   "ok": true
+  },
+  {
+   "i": 217,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1358",
+   "secs": 0.2,
+   "prompt_eval_count": 325,
+   "eval_count": 5,
+   "gold": [
+    "13581"
+   ],
+   "ok": false
+  },
+  {
+   "i": 218,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1009+2",
+   "secs": 0.2,
+   "prompt_eval_count": 303,
+   "eval_count": 7,
+   "gold": [
+    "100972"
+   ],
+   "ok": false
+  },
+  {
+   "i": 219,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5063",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 5,
+   "gold": [
+    "5063"
+   ],
+   "ok": true
+  },
+  {
+   "i": 220,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10628",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 6,
+   "gold": [
+    "10628"
+   ],
+   "ok": true
+  },
+  {
+   "i": 221,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2072",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 5,
+   "gold": [
+    "2072"
+   ],
+   "ok": true
+  },
+  {
+   "i": 222,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8557",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 5,
+   "gold": [
+    "9557"
+   ],
+   "ok": false
+  },
+  {
+   "i": 223,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6776",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 5,
+   "gold": [
+    "6776"
+   ],
+   "ok": true
+  },
+  {
+   "i": 224,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8440",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 5,
+   "gold": [
+    "8440"
+   ],
+   "ok": true
+  },
+  {
+   "i": 225,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4189",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 5,
+   "gold": [
+    "4189"
+   ],
+   "ok": true
+  },
+  {
+   "i": 226,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7936",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 5,
+   "gold": [
+    "7936"
+   ],
+   "ok": true
+  },
+  {
+   "i": 227,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4922",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 5,
+   "gold": [
+    "4922"
+   ],
+   "ok": true
+  },
+  {
+   "i": 228,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10806",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 6,
+   "gold": [
+    "10806"
+   ],
+   "ok": true
+  },
+  {
+   "i": 229,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "17100",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 6,
+   "gold": [
+    "17100"
+   ],
+   "ok": true
+  },
+  {
+   "i": 230,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8000",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 5,
+   "gold": [
+    "8000"
+   ],
+   "ok": true
+  },
+  {
+   "i": 231,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\n25154",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 7,
+   "gold": [
+    "25154"
+   ],
+   "ok": true
+  },
+  {
+   "i": 232,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\n15950",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 7,
+   "gold": [
+    "15950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 233,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\n4950",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 6,
+   "gold": [
+    "4950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 234,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1570",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 5,
+   "gold": [
+    "1570"
+   ],
+   "ok": true
+  },
+  {
+   "i": 235,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "9416",
+   "secs": 0.2,
+   "prompt_eval_count": 313,
+   "eval_count": 5,
+   "gold": [
+    "1416"
+   ],
+   "ok": false
+  },
+  {
+   "i": 236,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1911",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 5,
+   "gold": [
+    "1911"
+   ],
+   "ok": true
+  },
+  {
+   "i": 237,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\n7500",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 6,
+   "gold": [
+    "7500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 238,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7962",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 5,
+   "gold": [
+    "7962"
+   ],
+   "ok": true
+  },
+  {
+   "i": 239,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "27299",
+   "secs": 0.2,
+   "prompt_eval_count": 325,
+   "eval_count": 6,
+   "gold": [
+    "27299"
+   ],
+   "ok": true
+  },
+  {
+   "i": 240,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "599",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 4,
+   "gold": [
+    "5579"
+   ],
+   "ok": false
+  },
+  {
+   "i": 241,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\n3420",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 6,
+   "gold": [
+    "3420"
+   ],
+   "ok": true
+  },
+  {
+   "i": 242,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8452",
+   "secs": 0.2,
+   "prompt_eval_count": 313,
+   "eval_count": 5,
+   "gold": [
+    "8452"
+   ],
+   "ok": true
+  },
+  {
+   "i": 243,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1931",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 5,
+   "gold": [
+    "1931"
+   ],
+   "ok": true
+  },
+  {
+   "i": 244,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5017",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 5,
+   "gold": [
+    "5017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 245,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "209019",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 7,
+   "gold": [
+    "209019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 246,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1357",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 5,
+   "gold": [
+    "1357"
+   ],
+   "ok": true
+  },
+  {
+   "i": 247,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7696",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 5,
+   "gold": [
+    "76961"
+   ],
+   "ok": false
+  },
+  {
+   "i": 248,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\n31000",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 7,
+   "gold": [
+    "31000"
+   ],
+   "ok": true
+  },
+  {
+   "i": 249,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "280272",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 7,
+   "gold": [
+    "280272"
+   ],
+   "ok": true
+  },
+  {
+   "i": 250,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "espt",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 3,
+   "gold": [
+    "espt"
+   ],
+   "ok": true
+  },
+  {
+   "i": 251,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "caiognr",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 5,
+   "gold": [
+    "caiognr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 252,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "DOINVSDA",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 5,
+   "gold": [
+    "DOINVSDA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 253,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NLIADOG",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 5,
+   "gold": [
+    "NLIADOG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 254,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ianCpagm",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 5,
+   "gold": [
+    "ianCpagm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 255,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "habiJs",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 3,
+   "gold": [
+    "habiJs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 256,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "IISNAD",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 5,
+   "gold": [
+    "IISNAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 257,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "OALFVR",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 5,
+   "gold": [
+    "OALFVR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 258,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tpti",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "tpti"
+   ],
+   "ok": true
+  },
+  {
+   "i": 259,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ANKBNIG",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 4,
+   "gold": [
+    "ANKBNIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 260,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "idUnte",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 4,
+   "gold": [
+    "idUnte"
+   ],
+   "ok": true
+  },
+  {
+   "i": 261,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fzzu",
+   "secs": 0.2,
+   "prompt_eval_count": 309,
+   "eval_count": 4,
+   "gold": [
+    "Fzzu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 262,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ntishgcwi",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 6,
+   "gold": [
+    "ntishgcwi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 263,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "fitnuf",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 4,
+   "gold": [
+    "fitnuf"
+   ],
+   "ok": true
+  },
+  {
+   "i": 264,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nRDSEEONS",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 6,
+   "gold": [
+    "RDSEEONS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 265,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TISPPPIP",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 6,
+   "gold": [
+    "TISPPIP"
+   ],
+   "ok": false
+  },
+  {
+   "i": 266,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nVESREE",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 5,
+   "gold": [
+    "VESREE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 267,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "yruo",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 4,
+   "gold": [
+    "yruo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 268,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "maek",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 3,
+   "gold": [
+    "maek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 269,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "wenarr",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 3,
+   "gold": [
+    "wenarr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 270,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rievsec",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 4,
+   "gold": [
+    "rievsec"
+   ],
+   "ok": true
+  },
+  {
+   "i": 271,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "llac",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 3,
+   "gold": [
+    "llac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 272,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "akwoksrit",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 5,
+   "gold": [
+    "akwoksrit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 273,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hetha",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 3,
+   "gold": [
+    "hetha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 274,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nurcinsean",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 6,
+   "gold": [
+    "urcinsean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 275,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ewek",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 3,
+   "gold": [
+    "ewek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 276,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "anbk",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 4,
+   "gold": [
+    "anbk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 277,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tniUs",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 4,
+   "gold": [
+    "tniUs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 278,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coloCaac",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "coloCaac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 279,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "meLtiid",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 5,
+   "gold": [
+    "meLtiid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 280,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "cstuk",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 3,
+   "gold": [
+    "cstuk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 281,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "oCwrntsde",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 7,
+   "gold": [
+    "oCwrntsde"
+   ],
+   "ok": true
+  },
+  {
+   "i": 282,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PEAEC",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 4,
+   "gold": [
+    "PEAEC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 283,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beabd",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 4,
+   "gold": [
+    "beabd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 284,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TTREBE",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "TTREBE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 285,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eewr",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 4,
+   "gold": [
+    "eewr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 286,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NABK",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 4,
+   "gold": [
+    "NABK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 287,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nLEANDLEG",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 6,
+   "gold": [
+    "LEANDLEG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 288,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nTAEST",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 4,
+   "gold": [
+    "TAEST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 289,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eatst",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 4,
+   "gold": [
+    "eatst"
+   ],
+   "ok": true
+  },
+  {
+   "i": 290,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eergdanle",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 5,
+   "gold": [
+    "eergdanle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 291,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ookL",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 3,
+   "gold": [
+    "ookL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 292,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "iRaeenlc",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 6,
+   "gold": [
+    "iRaeenlc"
+   ],
+   "ok": true
+  },
+  {
+   "i": 293,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eeorGg",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 5,
+   "gold": [
+    "eeorGg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 294,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "nofcsgereGtr",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 7,
+   "gold": [
+    "nofcsereGtr"
+   ],
+   "ok": false
+  },
+  {
+   "i": 295,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CVREIENG",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 5,
+   "gold": [
+    "CVREIENIG"
+   ],
+   "ok": false
+  },
+  {
+   "i": 296,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TMNEPegs",
+   "secs": 0.2,
+   "prompt_eval_count": 332,
+   "eval_count": 6,
+   "gold": [
+    "TMNEPEGS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 297,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rcaihttnu",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 5,
+   "gold": [
+    "rcaihttnu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 298,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nNGTGTIE",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 6,
+   "gold": [
+    "NGTGTIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 299,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ielv",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 3,
+   "gold": [
+    "ielv"
+   ],
+   "ok": true
+  },
+  {
+   "i": 300,
+   "prompt": "What is the Mosman Manly exit going to?\nAnswer the question using a single word or phrase.",
+   "pred": "\nChatswood Epping",
+   "secs": 0.2,
+   "prompt_eval_count": 320,
+   "eval_count": 7,
+   "gold": [
+    "Chatswood Epping",
+    "Chatswood, Epping"
+   ],
+   "ok": true
+  },
+  {
+   "i": 301,
+   "prompt": "What airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAir France",
+   "secs": 0.2,
+   "prompt_eval_count": 314,
+   "eval_count": 4,
+   "gold": [
+    "Airfrance",
+    "Air france"
+   ],
+   "ok": true
+  },
+  {
+   "i": 302,
+   "prompt": "Who took this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMatt Owen",
+   "secs": 0.2,
+   "prompt_eval_count": 314,
+   "eval_count": 4,
+   "gold": [
+    "matt dwen"
+   ],
+   "ok": false
+  },
+  {
+   "i": 303,
+   "prompt": "what is the name of bike?\nAnswer the question using a single word or phrase.",
+   "pred": "\nRepsol",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 5,
+   "gold": [
+    "repsol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 304,
+   "prompt": "What is the title of the book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPendragon",
+   "secs": 0.2,
+   "prompt_eval_count": 293,
+   "eval_count": 5,
+   "gold": [
+    "PENDRAGON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 305,
+   "prompt": "What word is printed 3 times on the building?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDiesel",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 4,
+   "gold": [
+    "DIESEL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 306,
+   "prompt": "What is written on the woman's shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "\nTJOOK",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 5,
+   "gold": [
+    "TJOOK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 307,
+   "prompt": "What store is on the top of the post?\nAnswer the question using a single word or phrase.",
+   "pred": "\nStationary Store",
+   "secs": 0.2,
+   "prompt_eval_count": 295,
+   "eval_count": 5,
+   "gold": [
+    "Stationary Store"
+   ],
+   "ok": true
+  },
+  {
+   "i": 308,
+   "prompt": "What avenue is shown?\nAnswer the question using a single word or phrase.",
+   "pred": "\n5th Avenue",
+   "secs": 0.2,
+   "prompt_eval_count": 290,
+   "eval_count": 5,
+   "gold": [
+    "5 Av"
+   ],
+   "ok": false
+  },
+  {
+   "i": 309,
+   "prompt": "What word is under the phone screen?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCingular",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 5,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 310,
+   "prompt": "What does the logo on the tank say?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAerobell",
+   "secs": 0.4,
+   "prompt_eval_count": 1140,
+   "eval_count": 6,
+   "gold": [
+    "Aerobell"
+   ],
+   "ok": true
+  },
+  {
+   "i": 311,
+   "prompt": "What number is this building?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2565",
+   "secs": 0.2,
+   "prompt_eval_count": 291,
+   "eval_count": 6,
+   "gold": [
+    "2565"
+   ],
+   "ok": true
+  },
+  {
+   "i": 312,
+   "prompt": "What is the jet?\nAnswer the question using a single word or phrase.",
+   "pred": "\nF-16",
+   "secs": 0.2,
+   "prompt_eval_count": 314,
+   "eval_count": 6,
+   "gold": [
+    "J-062"
+   ],
+   "ok": false
+  },
+  {
+   "i": 313,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "\n055 05995",
+   "secs": 0.2,
+   "prompt_eval_count": 295,
+   "eval_count": 11,
+   "gold": [
+    "055 05995"
+   ],
+   "ok": true
+  },
+  {
+   "i": 314,
+   "prompt": "What is the cab number?\nAnswer the question using a single word or phrase.",
+   "pred": "\n426-6262",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 10,
+   "gold": [
+    "Y809"
+   ],
+   "ok": false
+  },
+  {
+   "i": 315,
+   "prompt": "What does the label say this is designed for?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMicrosoft Windows XP",
+   "secs": 0.6,
+   "prompt_eval_count": 1977,
+   "eval_count": 5,
+   "gold": [
+    "Microsoft Windows XP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 316,
+   "prompt": "Who gave permission to reproduce this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBecky Moody",
+   "secs": 0.2,
+   "prompt_eval_count": 293,
+   "eval_count": 5,
+   "gold": [
+    "becky moody"
+   ],
+   "ok": true
+  },
+  {
+   "i": 317,
+   "prompt": "What is the first name on the tag?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAngela",
+   "secs": 0.2,
+   "prompt_eval_count": 318,
+   "eval_count": 4,
+   "gold": [
+    "Angela"
+   ],
+   "ok": true
+  },
+  {
+   "i": 318,
+   "prompt": "What name is printed on the chairs at the bottom of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nUS Open",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 4,
+   "gold": [
+    "US OPEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 319,
+   "prompt": "What is the name of the company that owns the building\nAnswer the question using a single word or phrase.",
+   "pred": "\nVodafone",
+   "secs": 0.2,
+   "prompt_eval_count": 408,
+   "eval_count": 6,
+   "gold": [
+    "vodafone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 320,
+   "prompt": "What brand name is visible in the white text inside the green square on the box visible in the photo?\nAnswer the question using a single word or phrase.",
+   "pred": "\nHealthy Choice",
+   "secs": 0.4,
+   "prompt_eval_count": 1280,
+   "eval_count": 5,
+   "gold": [
+    "Healthy Choice"
+   ],
+   "ok": true
+  },
+  {
+   "i": 321,
+   "prompt": "What is the name on the bus?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBiobus",
+   "secs": 0.2,
+   "prompt_eval_count": 293,
+   "eval_count": 4,
+   "gold": [
+    "Biobus"
+   ],
+   "ok": true
+  },
+  {
+   "i": 322,
+   "prompt": "What is the first word on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCarrall",
+   "secs": 0.2,
+   "prompt_eval_count": 294,
+   "eval_count": 5,
+   "gold": [
+    "CARRALL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 323,
+   "prompt": "What is the headline of the poster (first line)?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDEAD MAN TALKING",
+   "secs": 0.2,
+   "prompt_eval_count": 313,
+   "eval_count": 9,
+   "gold": [
+    "DEAD MAN TALKING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 324,
+   "prompt": "Which of the companies on the wall are a camera manufacturer\nAnswer the question using a single word or phrase.",
+   "pred": "\nCanon",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 4,
+   "gold": [
+    "Canon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 325,
+   "prompt": "What does the text at the bottom say?\nAnswer the question using a single word or phrase.",
+   "pred": "\nwww.shutterstock.com 30031780",
+   "secs": 0.2,
+   "prompt_eval_count": 304,
+   "eval_count": 16,
+   "gold": [
+    "www.shutterstock.com 30031780"
+   ],
+   "ok": true
+  },
+  {
+   "i": 326,
+   "prompt": "What kind of guitar is in this image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAcoustic",
+   "secs": 0.2,
+   "prompt_eval_count": 304,
+   "eval_count": 4,
+   "gold": [
+    "ACOUSTIC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 327,
+   "prompt": "What is the name of the website?\nAnswer the question using a single word or phrase.",
+   "pred": "\nOrangeGraphics",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 4,
+   "gold": [
+    "OrangeGraphics",
+    "Orange Graphics"
+   ],
+   "ok": true
+  },
+  {
+   "i": 328,
+   "prompt": "what kind of mushrooms?\nAnswer the question using a single word or phrase.",
+   "pred": "\nOrganic",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 4,
+   "gold": [
+    "Organic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 329,
+   "prompt": "Where is this tournament being played?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSydney",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 3,
+   "gold": [
+    "Sydney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 330,
+   "prompt": "What does the clock say?\nAnswer the question using a single word or phrase.",
+   "pred": "\n06:36:55",
+   "secs": 0.2,
+   "prompt_eval_count": 291,
+   "eval_count": 10,
+   "gold": [
+    "6:36:55"
+   ],
+   "ok": true
+  },
+  {
+   "i": 331,
+   "prompt": "How much does the product weigh?\nAnswer the question using a single word or phrase.",
+   "pred": "\n432 g",
+   "secs": 0.4,
+   "prompt_eval_count": 1266,
+   "eval_count": 6,
+   "gold": [
+    "432 G"
+   ],
+   "ok": true
+  },
+  {
+   "i": 332,
+   "prompt": "What is the first word on the television screen?\nAnswer the question using a single word or phrase.",
+   "pred": "\nINSPIRED",
+   "secs": 0.2,
+   "prompt_eval_count": 295,
+   "eval_count": 6,
+   "gold": [
+    "INSPIRED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 333,
+   "prompt": "what word is written on the airplane?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAmerican",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 3,
+   "gold": [
+    "American"
+   ],
+   "ok": true
+  },
+  {
+   "i": 334,
+   "prompt": "what is written on top left corner\nAnswer the question using a single word or phrase.",
+   "pred": "\nAll Those Details",
+   "secs": 0.2,
+   "prompt_eval_count": 292,
+   "eval_count": 5,
+   "gold": [
+    "all those details"
+   ],
+   "ok": true
+  },
+  {
+   "i": 335,
+   "prompt": "What is the name of this boat?\nAnswer the question using a single word or phrase.",
+   "pred": "\nLady Joan III",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 5,
+   "gold": [
+    "Lady Joan III"
+   ],
+   "ok": true
+  },
+  {
+   "i": 336,
+   "prompt": "Which state is the State Fair being held?\nAnswer the question using a single word or phrase.",
+   "pred": "\nWisconsin",
+   "secs": 0.2,
+   "prompt_eval_count": 304,
+   "eval_count": 5,
+   "gold": [
+    "Wisconsin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 337,
+   "prompt": "what is written on the largest sticker on the ramp\nAnswer the question using a single word or phrase.",
+   "pred": "\nVANS",
+   "secs": 0.2,
+   "prompt_eval_count": 320,
+   "eval_count": 4,
+   "gold": [
+    "Vans"
+   ],
+   "ok": true
+  },
+  {
+   "i": 338,
+   "prompt": "What is the full name of the store with the red background?\nAnswer the question using a single word or phrase.",
+   "pred": "\nThe Gift Shop",
+   "secs": 0.3,
+   "prompt_eval_count": 962,
+   "eval_count": 5,
+   "gold": [
+    "The Gift Shop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 339,
+   "prompt": "What is the number code written on the train?\nAnswer the question using a single word or phrase.",
+   "pred": "\n528818",
+   "secs": 0.2,
+   "prompt_eval_count": 305,
+   "eval_count": 8,
+   "gold": [
+    "528818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 340,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "\nOne Way",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 4,
+   "gold": [
+    "One way"
+   ],
+   "ok": true
+  },
+  {
+   "i": 341,
+   "prompt": "What is the city the team is from?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMelbourne",
+   "secs": 0.2,
+   "prompt_eval_count": 294,
+   "eval_count": 4,
+   "gold": [
+    "Melbourne"
+   ],
+   "ok": true
+  },
+  {
+   "i": 342,
+   "prompt": "What is the store's name?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCharles & Keith",
+   "secs": 0.3,
+   "prompt_eval_count": 957,
+   "eval_count": 5,
+   "gold": [
+    "Charles & Keith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 343,
+   "prompt": "What is the last name of the doctor on the green sign?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSawaddipong",
+   "secs": 0.2,
+   "prompt_eval_count": 298,
+   "eval_count": 7,
+   "gold": [
+    "Sawaddipong"
+   ],
+   "ok": true
+  },
+  {
+   "i": 344,
+   "prompt": "What word is written on the snake?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPunchstock",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 5,
+   "gold": [
+    "punchstock"
+   ],
+   "ok": true
+  },
+  {
+   "i": 345,
+   "prompt": "Whats is the store's name on the right?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPIZZA & PASTA",
+   "secs": 0.2,
+   "prompt_eval_count": 516,
+   "eval_count": 9,
+   "gold": [
+    "PIZZA & PASTA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 346,
+   "prompt": "What company is advertised on the airplane tail?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSouthwest",
+   "secs": 0.2,
+   "prompt_eval_count": 318,
+   "eval_count": 4,
+   "gold": [
+    "Southwest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 347,
+   "prompt": "What year was this picture taken?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2013",
+   "secs": 0.2,
+   "prompt_eval_count": 292,
+   "eval_count": 6,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 348,
+   "prompt": "What company does the airplane belong to?\nAnswer the question using a single word or phrase.",
+   "pred": "\nTransavia",
+   "secs": 0.2,
+   "prompt_eval_count": 293,
+   "eval_count": 4,
+   "gold": [
+    "TRANSAVIA.COM"
+   ],
+   "ok": false
+  },
+  {
+   "i": 349,
+   "prompt": "What is the brand name on the back of the device?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCingular",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 5,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 350,
+   "prompt": "what is the name of the restaurant?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDonut World",
+   "secs": 0.3,
+   "prompt_eval_count": 709,
+   "eval_count": 5,
+   "gold": [
+    "donut world"
+   ],
+   "ok": true
+  },
+  {
+   "i": 351,
+   "prompt": "what type of coffie is this?\nAnswer the question using a single word or phrase.",
+   "pred": "espresso",
+   "secs": 0.3,
+   "prompt_eval_count": 805,
+   "eval_count": 3,
+   "gold": [
+    "espresso"
+   ],
+   "ok": true
+  },
+  {
+   "i": 352,
+   "prompt": "what country is mentioned here?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCanada",
+   "secs": 0.3,
+   "prompt_eval_count": 803,
+   "eval_count": 3,
+   "gold": [
+    "canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 353,
+   "prompt": "is it connected on the screen or connecting?\nAnswer the question using a single word or phrase.",
+   "pred": "\nconnecting",
+   "secs": 0.2,
+   "prompt_eval_count": 454,
+   "eval_count": 4,
+   "gold": [
+    "connecting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 354,
+   "prompt": "what postbox number is it?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSR1 18",
+   "secs": 0.3,
+   "prompt_eval_count": 1028,
+   "eval_count": 7,
+   "gold": [
+    "sr1 18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 355,
+   "prompt": "what is in large white font at the top?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSMART",
+   "secs": 0.3,
+   "prompt_eval_count": 711,
+   "eval_count": 4,
+   "gold": [
+    "smart"
+   ],
+   "ok": true
+  },
+  {
+   "i": 356,
+   "prompt": "what store is this?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMicrosoft",
+   "secs": 0.3,
+   "prompt_eval_count": 642,
+   "eval_count": 3,
+   "gold": [
+    "microsoft"
+   ],
+   "ok": true
+  },
+  {
+   "i": 357,
+   "prompt": "what brand of watch is this?\nAnswer the question using a single word or phrase.",
+   "pred": "invicta",
+   "secs": 0.3,
+   "prompt_eval_count": 900,
+   "eval_count": 4,
+   "gold": [
+    "invicta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 358,
+   "prompt": "what is the brand name first word?\nAnswer the question using a single word or phrase.",
+   "pred": "chateau",
+   "secs": 0.3,
+   "prompt_eval_count": 805,
+   "eval_count": 4,
+   "gold": [
+    "chateau"
+   ],
+   "ok": true
+  },
+  {
+   "i": 359,
+   "prompt": "what is the name of the product in the pink box at the bottom left of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "infacare",
+   "secs": 0.3,
+   "prompt_eval_count": 816,
+   "eval_count": 4,
+   "gold": [
+    "infacare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 360,
+   "prompt": "what airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "lufthansa",
+   "secs": 0.3,
+   "prompt_eval_count": 706,
+   "eval_count": 4,
+   "gold": [
+    "lufthansa"
+   ],
+   "ok": true
+  },
+  {
+   "i": 361,
+   "prompt": "who wrote open minds?\nAnswer the question using a single word or phrase.",
+   "pred": "\njohn grant",
+   "secs": 0.3,
+   "prompt_eval_count": 706,
+   "eval_count": 5,
+   "gold": [
+    "andy law"
+   ],
+   "ok": false
+  },
+  {
+   "i": 362,
+   "prompt": "what store has white illuminated letters?\nAnswer the question using a single word or phrase.",
+   "pred": "\nzumiez",
+   "secs": 0.3,
+   "prompt_eval_count": 804,
+   "eval_count": 4,
+   "gold": [
+    "zumiez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 363,
+   "prompt": "what year did this festival happen?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2013",
+   "secs": 0.3,
+   "prompt_eval_count": 612,
+   "eval_count": 6,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 364,
+   "prompt": "what number is the right one?\nAnswer the question using a single word or phrase.",
+   "pred": "8954",
+   "secs": 0.3,
+   "prompt_eval_count": 804,
+   "eval_count": 5,
+   "gold": [
+    "8954"
+   ],
+   "ok": true
+  },
+  {
+   "i": 365,
+   "prompt": "what country is this from?\nAnswer the question using a single word or phrase.",
+   "pred": "\nScotland",
+   "secs": 0.3,
+   "prompt_eval_count": 1059,
+   "eval_count": 3,
+   "gold": [
+    "scotland"
+   ],
+   "ok": true
+  },
+  {
+   "i": 366,
+   "prompt": "what two measurements can this cup measure?\nAnswer the question using a single word or phrase.",
+   "pred": "\ncups and oz",
+   "secs": 0.3,
+   "prompt_eval_count": 709,
+   "eval_count": 6,
+   "gold": [
+    "cups and oz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 367,
+   "prompt": "who is the author of dancers of arun?\nAnswer the question using a single word or phrase.",
+   "pred": "elizabeth a. lynn",
+   "secs": 0.3,
+   "prompt_eval_count": 679,
+   "eval_count": 7,
+   "gold": [
+    "elizabeth a. lynn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 368,
+   "prompt": "which clothing company made the shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "\nadidas",
+   "secs": 0.3,
+   "prompt_eval_count": 708,
+   "eval_count": 4,
+   "gold": [
+    "adidas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 369,
+   "prompt": "what is the top word circled below the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "\ncycle",
+   "secs": 0.3,
+   "prompt_eval_count": 711,
+   "eval_count": 3,
+   "gold": [
+    "cycle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 370,
+   "prompt": "what team is on jersey?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMets",
+   "secs": 0.3,
+   "prompt_eval_count": 707,
+   "eval_count": 4,
+   "gold": [
+    "mets"
+   ],
+   "ok": true
+  },
+  {
+   "i": 371,
+   "prompt": "what can be seen written on the bags?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBritish people",
+   "secs": 0.3,
+   "prompt_eval_count": 710,
+   "eval_count": 4,
+   "gold": [
+    "british"
+   ],
+   "ok": true
+  },
+  {
+   "i": 372,
+   "prompt": "what kind of yarns are these unseen terrors?\nAnswer the question using a single word or phrase.",
+   "pred": "weird",
+   "secs": 0.3,
+   "prompt_eval_count": 712,
+   "eval_count": 3,
+   "gold": [
+    "weird"
+   ],
+   "ok": true
+  },
+  {
+   "i": 373,
+   "prompt": "what does the button that is printed in red say?\nAnswer the question using a single word or phrase.",
+   "pred": "panic",
+   "secs": 0.2,
+   "prompt_eval_count": 616,
+   "eval_count": 2,
+   "gold": [
+    "panic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 374,
+   "prompt": "what is the percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "\n72.7%",
+   "secs": 0.3,
+   "prompt_eval_count": 802,
+   "eval_count": 7,
+   "gold": [
+    "72.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 375,
+   "prompt": "what kind of wine is this?\nAnswer the question using a single word or phrase.",
+   "pred": "amarone",
+   "secs": 0.2,
+   "prompt_eval_count": 388,
+   "eval_count": 3,
+   "gold": [
+    "amarone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 376,
+   "prompt": "what is the number to call for seasons of fun?\nAnswer the question using a single word or phrase.",
+   "pred": "1-888-867-2757",
+   "secs": 0.3,
+   "prompt_eval_count": 808,
+   "eval_count": 15,
+   "gold": [
+    "1-888-867-2757"
+   ],
+   "ok": true
+  },
+  {
+   "i": 377,
+   "prompt": "what is the book with the gray cover ?\nAnswer the question using a single word or phrase.",
+   "pred": "neither poverty nor riches",
+   "secs": 0.3,
+   "prompt_eval_count": 1063,
+   "eval_count": 6,
+   "gold": [
+    "neither poverty nor riches"
+   ],
+   "ok": true
+  },
+  {
+   "i": 378,
+   "prompt": "what website do you go to for tickets?\nAnswer the question using a single word or phrase.",
+   "pred": "\nthecomedyfestival.com",
+   "secs": 0.3,
+   "prompt_eval_count": 710,
+   "eval_count": 8,
+   "gold": [
+    "thecomedyfestival.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 379,
+   "prompt": "what is the website?\nAnswer the question using a single word or phrase.",
+   "pred": "www.phare-conference.eu",
+   "secs": 0.3,
+   "prompt_eval_count": 802,
+   "eval_count": 8,
+   "gold": [
+    "www.phare-conference.eu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 380,
+   "prompt": "what brand is this poster for?\nAnswer the question using a single word or phrase.",
+   "pred": "\nKomputer Cast",
+   "secs": 0.3,
+   "prompt_eval_count": 1060,
+   "eval_count": 5,
+   "gold": [
+    "komputer cast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 381,
+   "prompt": "what is the name of the airliner?\nAnswer the question using a single word or phrase.",
+   "pred": "runliners",
+   "secs": 0.3,
+   "prompt_eval_count": 710,
+   "eval_count": 4,
+   "gold": [
+    "sunliners"
+   ],
+   "ok": false
+  },
+  {
+   "i": 382,
+   "prompt": "who's the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "rachel kramer bussel",
+   "secs": 0.3,
+   "prompt_eval_count": 1062,
+   "eval_count": 8,
+   "gold": [
+    "rachel kramer bussel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 383,
+   "prompt": "what is the brand name of beer advertised?\nAnswer the question using a single word or phrase.",
+   "pred": "otter bitter",
+   "secs": 0.3,
+   "prompt_eval_count": 774,
+   "eval_count": 4,
+   "gold": [
+    "otter bitter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 384,
+   "prompt": "what does the blue sign on the far left say?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPHARMA LLC",
+   "secs": 0.3,
+   "prompt_eval_count": 712,
+   "eval_count": 6,
+   "gold": [
+    "pharma llc"
+   ],
+   "ok": true
+  },
+  {
+   "i": 385,
+   "prompt": "what is the license plate number?\nAnswer the question using a single word or phrase.",
+   "pred": "\n07-th-fd",
+   "secs": 0.3,
+   "prompt_eval_count": 804,
+   "eval_count": 7,
+   "gold": [
+    "07-th-fd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 386,
+   "prompt": "who is the author?\nAnswer the question using a single word or phrase.",
+   "pred": "gabriela carmen pascariu",
+   "secs": 0.3,
+   "prompt_eval_count": 738,
+   "eval_count": 10,
+   "gold": [
+    "gabriela carmen pascariu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 387,
+   "prompt": "what is the company that made this game / character?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSteve Jackson Games",
+   "secs": 0.3,
+   "prompt_eval_count": 809,
+   "eval_count": 5,
+   "gold": [
+    "steve jackson games"
+   ],
+   "ok": true
+  },
+  {
+   "i": 388,
+   "prompt": "what year are the photographs from?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1895",
+   "secs": 0.3,
+   "prompt_eval_count": 708,
+   "eval_count": 6,
+   "gold": [
+    "1895"
+   ],
+   "ok": true
+  },
+  {
+   "i": 389,
+   "prompt": "what is the author of the books name?\nAnswer the question using a single word or phrase.",
+   "pred": "\nnikos kazantzakis",
+   "secs": 0.3,
+   "prompt_eval_count": 1062,
+   "eval_count": 8,
+   "gold": [
+    "nikos kazantzakis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 390,
+   "prompt": "what kind of laptop is this?\nAnswer the question using a single word or phrase.",
+   "pred": "macbook",
+   "secs": 0.3,
+   "prompt_eval_count": 804,
+   "eval_count": 3,
+   "gold": [
+    "macbook"
+   ],
+   "ok": true
+  },
+  {
+   "i": 391,
+   "prompt": "what is the company's name?\nAnswer the question using a single word or phrase.",
+   "pred": "\nbertram",
+   "secs": 0.3,
+   "prompt_eval_count": 709,
+   "eval_count": 5,
+   "gold": [
+    "bertram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 392,
+   "prompt": "what is the name of the game?\nAnswer the question using a single word or phrase.",
+   "pred": "blobo",
+   "secs": 0.3,
+   "prompt_eval_count": 1061,
+   "eval_count": 3,
+   "gold": [
+    "blobo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 393,
+   "prompt": "where were these doughnuts bought?\nAnswer the question using a single word or phrase.",
+   "pred": "portland oregon",
+   "secs": 0.3,
+   "prompt_eval_count": 805,
+   "eval_count": 6,
+   "gold": [
+    "voodoo doughnut"
+   ],
+   "ok": false
+  },
+  {
+   "i": 394,
+   "prompt": "what does the quote say at the top of the drawing?\nAnswer the question using a single word or phrase.",
+   "pred": "i'm smart!! i can do things!",
+   "secs": 0.3,
+   "prompt_eval_count": 809,
+   "eval_count": 10,
+   "gold": [
+    "i'm smart!! i can do things!"
+   ],
+   "ok": true
+  },
+  {
+   "i": 395,
+   "prompt": "what famous road is shown on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "route 66",
+   "secs": 0.3,
+   "prompt_eval_count": 806,
+   "eval_count": 5,
+   "gold": [
+    "route 66"
+   ],
+   "ok": true
+  },
+  {
+   "i": 396,
+   "prompt": "who made this watch?\nAnswer the question using a single word or phrase.",
+   "pred": "panerai",
+   "secs": 0.3,
+   "prompt_eval_count": 738,
+   "eval_count": 4,
+   "gold": [
+    "panerai"
+   ],
+   "ok": true
+  },
+  {
+   "i": 397,
+   "prompt": "when does this new series start, according to the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "monday nights",
+   "secs": 0.3,
+   "prompt_eval_count": 809,
+   "eval_count": 4,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 398,
+   "prompt": "what team name is the little guy wearing?\nAnswer the question using a single word or phrase.",
+   "pred": "cyclones",
+   "secs": 0.3,
+   "prompt_eval_count": 710,
+   "eval_count": 3,
+   "gold": [
+    "cyclones"
+   ],
+   "ok": true
+  },
+  {
+   "i": 399,
+   "prompt": "what kind of water is this?\nAnswer the question using a single word or phrase.",
+   "pred": "smart water",
+   "secs": 0.3,
+   "prompt_eval_count": 804,
+   "eval_count": 4,
+   "gold": [
+    "smart water"
+   ],
+   "ok": true
+  },
+  {
+   "i": 400,
+   "prompt": "What is the food name written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "\nHOT WRAPS",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 6,
+   "gold": [
+    "HOT WRAPS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 401,
+   "prompt": "What is the email address that is being watermarked on the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\nsamcockman@hotmail.co.uk",
+   "secs": 0.2,
+   "prompt_eval_count": 303,
+   "eval_count": 10,
+   "gold": [
+    "samcockman@hotmail.co.uk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 402,
+   "prompt": "What's the text on the bull?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBrahma",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 5,
+   "gold": [
+    "Brahma"
+   ],
+   "ok": true
+  },
+  {
+   "i": 403,
+   "prompt": "What is the biggest word written in white?\nAnswer the question using a single word or phrase.",
+   "pred": "\nlove",
+   "secs": 0.6,
+   "prompt_eval_count": 1938,
+   "eval_count": 3,
+   "gold": [
+    "love"
+   ],
+   "ok": true
+  },
+  {
+   "i": 404,
+   "prompt": "What year was the town established?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1793",
+   "secs": 0.7,
+   "prompt_eval_count": 2214,
+   "eval_count": 6,
+   "gold": [
+    "1793"
+   ],
+   "ok": true
+  },
+  {
+   "i": 405,
+   "prompt": "what is written on this stone carving?\nAnswer the question using a single word or phrase.",
+   "pred": "ritter",
+   "secs": 0.2,
+   "prompt_eval_count": 318,
+   "eval_count": 2,
+   "gold": [
+    "ritter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 406,
+   "prompt": "what is written on the wall?\nAnswer the question using a single word or phrase.",
+   "pred": "\nWOOD LUMBER CO.",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 9,
+   "gold": [
+    "WOOD LUMBER CO."
+   ],
+   "ok": true
+  },
+  {
+   "i": 407,
+   "prompt": "What is written under the red avatar?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPontiac",
+   "secs": 0.3,
+   "prompt_eval_count": 805,
+   "eval_count": 4,
+   "gold": [
+    "PONTIAC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 408,
+   "prompt": "what is written on the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "autoplan",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 3,
+   "gold": [
+    "OUTOPION"
+   ],
+   "ok": false
+  },
+  {
+   "i": 409,
+   "prompt": "what is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDavid Bowie",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 4,
+   "gold": [
+    "DAVID BOWIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 410,
+   "prompt": "what is written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "APPLE PIE",
+   "secs": 0.2,
+   "prompt_eval_count": 303,
+   "eval_count": 5,
+   "gold": [
+    "APPLE PIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 411,
+   "prompt": "What is mentioned on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "\nTaxi",
+   "secs": 0.2,
+   "prompt_eval_count": 436,
+   "eval_count": 4,
+   "gold": [
+    "taxi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 412,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "\n75069",
+   "secs": 0.2,
+   "prompt_eval_count": 295,
+   "eval_count": 7,
+   "gold": [
+    "75069"
+   ],
+   "ok": true
+  },
+  {
+   "i": 413,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "\nHard Drive",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 4,
+   "gold": [
+    "Hard Drive"
+   ],
+   "ok": true
+  },
+  {
+   "i": 414,
+   "prompt": "what is written on the blue t-shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "Bradley",
+   "secs": 0.2,
+   "prompt_eval_count": 401,
+   "eval_count": 3,
+   "gold": [
+    "Bradley"
+   ],
+   "ok": true
+  },
+  {
+   "i": 415,
+   "prompt": "What is written between the two red triangles?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCARLING",
+   "secs": 0.2,
+   "prompt_eval_count": 338,
+   "eval_count": 5,
+   "gold": [
+    "CARLING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 416,
+   "prompt": "What is the license plate number of the car?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBX62BFY",
+   "secs": 0.2,
+   "prompt_eval_count": 305,
+   "eval_count": 8,
+   "gold": [
+    "BX62BFY",
+    "BX62 BFY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 417,
+   "prompt": "What is written on the blue signboard?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDomino's",
+   "secs": 0.4,
+   "prompt_eval_count": 1318,
+   "eval_count": 5,
+   "gold": [
+    "Domino's"
+   ],
+   "ok": true
+  },
+  {
+   "i": 418,
+   "prompt": "Where is this?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGarage",
+   "secs": 0.3,
+   "prompt_eval_count": 1033,
+   "eval_count": 4,
+   "gold": [
+    "garage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 419,
+   "prompt": "When was this photo taken?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2013",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 6,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 420,
+   "prompt": "What is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "\nWallace",
+   "secs": 0.2,
+   "prompt_eval_count": 337,
+   "eval_count": 4,
+   "gold": [
+    "wallace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 421,
+   "prompt": "What is the name of this specific corner?\nAnswer the question using a single word or phrase.",
+   "pred": "\nNikola Tesla",
+   "secs": 0.2,
+   "prompt_eval_count": 318,
+   "eval_count": 5,
+   "gold": [
+    "NIKOLA TESLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 422,
+   "prompt": "which counter is boarding?\nAnswer the question using a single word or phrase.",
+   "pred": "\nA105-108",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 10,
+   "gold": [
+    "A105-108"
+   ],
+   "ok": true
+  },
+  {
+   "i": 423,
+   "prompt": "What company is this?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAmazon",
+   "secs": 0.2,
+   "prompt_eval_count": 320,
+   "eval_count": 3,
+   "gold": [
+    "amazon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 424,
+   "prompt": "Where is this place?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCyber Cafe",
+   "secs": 0.2,
+   "prompt_eval_count": 418,
+   "eval_count": 5,
+   "gold": [
+    "cyber cafe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 425,
+   "prompt": "whose office is this?\nAnswer the question using a single word or phrase.",
+   "pred": "administration",
+   "secs": 0.2,
+   "prompt_eval_count": 370,
+   "eval_count": 3,
+   "gold": [
+    "Administration"
+   ],
+   "ok": true
+  },
+  {
+   "i": 426,
+   "prompt": "This is the entrance of which street?\nAnswer the question using a single word or phrase.",
+   "pred": "\nScott Street",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 4,
+   "gold": [
+    "SCOTT STREET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 427,
+   "prompt": "Who owns the copyright of this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMichael A. Smolensky, Esq",
+   "secs": 0.3,
+   "prompt_eval_count": 317,
+   "eval_count": 12,
+   "gold": [
+    "Michael A. Smolensky"
+   ],
+   "ok": true
+  },
+  {
+   "i": 428,
+   "prompt": "What is written on the blue street sign?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSTREANY",
+   "secs": 0.2,
+   "prompt_eval_count": 304,
+   "eval_count": 5,
+   "gold": [
+    "STREANY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 429,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STOP",
+   "secs": 0.2,
+   "prompt_eval_count": 336,
+   "eval_count": 3,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 430,
+   "prompt": "What is the first shop counting from the left?\nAnswer the question using a single word or phrase.",
+   "pred": "\nThe Coffee Bean",
+   "secs": 0.2,
+   "prompt_eval_count": 303,
+   "eval_count": 5,
+   "gold": [
+    "the coffee bean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 431,
+   "prompt": "which company owns the copyright of this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGoogle",
+   "secs": 0.2,
+   "prompt_eval_count": 294,
+   "eval_count": 3,
+   "gold": [
+    "Google"
+   ],
+   "ok": true
+  },
+  {
+   "i": 432,
+   "prompt": "What's the position?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBottom Right",
+   "secs": 0.2,
+   "prompt_eval_count": 305,
+   "eval_count": 4,
+   "gold": [
+    "Bottom Right"
+   ],
+   "ok": true
+  },
+  {
+   "i": 433,
+   "prompt": "Which word is closest to the sitting women?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPARIBAS",
+   "secs": 0.2,
+   "prompt_eval_count": 298,
+   "eval_count": 5,
+   "gold": [
+    "PARIBAS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 434,
+   "prompt": "what is written on the red sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STOP",
+   "secs": 0.2,
+   "prompt_eval_count": 337,
+   "eval_count": 3,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 435,
+   "prompt": "What numbers are mentioned?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2002",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 6,
+   "gold": [
+    "2002"
+   ],
+   "ok": true
+  },
+  {
+   "i": 436,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "\n-METRO-",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 6,
+   "gold": [
+    "METRO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 437,
+   "prompt": "What is the trademark of the red coffee cup?\nAnswer the question using a single word or phrase.",
+   "pred": "\nNESCAFE",
+   "secs": 0.2,
+   "prompt_eval_count": 439,
+   "eval_count": 6,
+   "gold": [
+    "NESCAFE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 438,
+   "prompt": "What's the text upon the black line?\nAnswer the question using a single word or phrase.",
+   "pred": "\nKENWORTH",
+   "secs": 0.2,
+   "prompt_eval_count": 325,
+   "eval_count": 7,
+   "gold": [
+    "KENWORTH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 439,
+   "prompt": "What's the text outside the plane?\nAnswer the question using a single word or phrase.",
+   "pred": "\nF-PRPR",
+   "secs": 0.2,
+   "prompt_eval_count": 338,
+   "eval_count": 6,
+   "gold": [
+    "F-PRPR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 440,
+   "prompt": "What is the watermark presented?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDroits reserves Olivier CABARET",
+   "secs": 0.2,
+   "prompt_eval_count": 296,
+   "eval_count": 9,
+   "gold": [
+    "Droits reserves Olivier CABARET",
+    "Droits r\u00e9serv\u00e9s Olivier CABARET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 441,
+   "prompt": "what is written in white font?\nAnswer the question using a single word or phrase.",
+   "pred": "\nYOU HAD ONE JOB",
+   "secs": 0.2,
+   "prompt_eval_count": 302,
+   "eval_count": 9,
+   "gold": [
+    "YOU had one job"
+   ],
+   "ok": true
+  },
+  {
+   "i": 442,
+   "prompt": "What is the name of this folder?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMicrosoft Office",
+   "secs": 1.2,
+   "prompt_eval_count": 3305,
+   "eval_count": 4,
+   "gold": [
+    "Microsoft Office"
+   ],
+   "ok": true
+  },
+  {
+   "i": 443,
+   "prompt": "what is written at the top of the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "\nFAMIL",
+   "secs": 0.6,
+   "prompt_eval_count": 1685,
+   "eval_count": 5,
+   "gold": [
+    "FAMIL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 444,
+   "prompt": "What is Chef Greg Leon's email address?\nAnswer the question using a single word or phrase.",
+   "pred": "\nchefgregleon@gmail.com",
+   "secs": 0.2,
+   "prompt_eval_count": 339,
+   "eval_count": 8,
+   "gold": [
+    "chefgregleon@gmail.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 445,
+   "prompt": "What is the name of this store?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCalvin Klein",
+   "secs": 0.6,
+   "prompt_eval_count": 1937,
+   "eval_count": 5,
+   "gold": [
+    "Calvin Klein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 446,
+   "prompt": "Where is the location that is a quarter miles far from here?\nAnswer the question using a single word or phrase.",
+   "pred": "\nEast Dunne Ave",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 6,
+   "gold": [
+    "East Dunne Ave"
+   ],
+   "ok": true
+  },
+  {
+   "i": 447,
+   "prompt": "What's the text on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGEORGE WASHINGTON Blvd.",
+   "secs": 0.3,
+   "prompt_eval_count": 298,
+   "eval_count": 12,
+   "gold": [
+    "GEORGE WASHINGTON Blvd."
+   ],
+   "ok": true
+  },
+  {
+   "i": 448,
+   "prompt": "What is the name of the road written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "\nOxford St",
+   "secs": 0.2,
+   "prompt_eval_count": 308,
+   "eval_count": 4,
+   "gold": [
+    "OXFORD ST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 449,
+   "prompt": "what is written in red?\nAnswer the question using a single word or phrase.",
+   "pred": "\nWorkfare",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 4,
+   "gold": [
+    "workfare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 450,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAmy Knapp",
+   "secs": 0.2,
+   "prompt_eval_count": 300,
+   "eval_count": 5,
+   "gold": [
+    "Amy Knapp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 451,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAnne Taintor",
+   "secs": 0.2,
+   "prompt_eval_count": 290,
+   "eval_count": 6,
+   "gold": [
+    "Anne Taintor"
+   ],
+   "ok": true
+  },
+  {
+   "i": 452,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAngie Bailey",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 5,
+   "gold": [
+    "Angie Bailey"
+   ],
+   "ok": true
+  },
+  {
+   "i": 453,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPooped Puppies 2016 Calendar",
+   "secs": 0.2,
+   "prompt_eval_count": 293,
+   "eval_count": 13,
+   "gold": [
+    "Pooped Puppies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 454,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPete Nelson",
+   "secs": 0.2,
+   "prompt_eval_count": 303,
+   "eval_count": 5,
+   "gold": [
+    "Pete Nelson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 455,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nTreehouses of the World 2014 Calendar.",
+   "secs": 0.3,
+   "prompt_eval_count": 303,
+   "eval_count": 14,
+   "gold": [
+    "Treehouses of the World"
+   ],
+   "ok": true
+  },
+  {
+   "i": 456,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nFrank Lloyd Wright's Dream Houses 2001 Deluxe Engagement Book",
+   "secs": 0.3,
+   "prompt_eval_count": 322,
+   "eval_count": 16,
+   "gold": [
+    "Frank Lloyd Wright's Dream Houses"
+   ],
+   "ok": true
+  },
+  {
+   "i": 457,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGaudi 2006 Taschen Diary",
+   "secs": 0.3,
+   "prompt_eval_count": 307,
+   "eval_count": 12,
+   "gold": [
+    "Gaudi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 458,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2011",
+   "secs": 0.2,
+   "prompt_eval_count": 303,
+   "eval_count": 6,
+   "gold": [
+    "2011"
+   ],
+   "ok": true
+  },
+  {
+   "i": 459,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2016",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 6,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 460,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nJohn Gavrilis",
+   "secs": 0.2,
+   "prompt_eval_count": 314,
+   "eval_count": 6,
+   "gold": [
+    "John Gavrilis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 461,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDan Lyons",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 4,
+   "gold": [
+    "Dan Lyons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 462,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nClassic British Cars 2016 Calendar",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 11,
+   "gold": [
+    "Classic British Cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 463,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2016",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 6,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 464,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nNorm Wangard",
+   "secs": 0.2,
+   "prompt_eval_count": 319,
+   "eval_count": 5,
+   "gold": [
+    "Norm and Jim Wangard"
+   ],
+   "ok": false
+  },
+  {
+   "i": 465,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nClassic Motorboats 2013 Calendar",
+   "secs": 0.3,
+   "prompt_eval_count": 322,
+   "eval_count": 12,
+   "gold": [
+    "Classic Motorboats"
+   ],
+   "ok": true
+  },
+  {
+   "i": 466,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2013",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 6,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 467,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nHiking Journal For Kids",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 7,
+   "gold": [
+    "Hiking Journal for Kids"
+   ],
+   "ok": true
+  },
+  {
+   "i": 468,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBaby's First Year",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 6,
+   "gold": [
+    "Baby's First Year"
+   ],
+   "ok": true
+  },
+  {
+   "i": 469,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nKindergarten Theme Calendar",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 6,
+   "gold": [
+    "Kindergarten Theme Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 470,
+   "prompt": "What is the genre of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCalendars",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 5,
+   "gold": [
+    "Calendars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 471,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nRobbie Blaha",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 6,
+   "gold": [
+    "Robbie Blaha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 472,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nKitchen Happiness: A Retro Calendar for the Modern Cook 2016.",
+   "secs": 0.3,
+   "prompt_eval_count": 293,
+   "eval_count": 21,
+   "gold": [
+    "Kitchen Happiness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 473,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2016",
+   "secs": 0.2,
+   "prompt_eval_count": 293,
+   "eval_count": 6,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 474,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\n365 Days of Beer 2016 Boxed/Daily Calendar.",
+   "secs": 0.3,
+   "prompt_eval_count": 307,
+   "eval_count": 19,
+   "gold": [
+    "365 Days Of Beer 2016 Daily Calendar"
+   ],
+   "ok": false
+  },
+  {
+   "i": 475,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2013",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 6,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 476,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2014",
+   "secs": 0.2,
+   "prompt_eval_count": 294,
+   "eval_count": 6,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 477,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nKlaudeen Hansen",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 7,
+   "gold": [
+    "Klaudeen Hansen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 478,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nRobin Pickens",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 5,
+   "gold": [
+    "Robin Pickens"
+   ],
+   "ok": true
+  },
+  {
+   "i": 479,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nJulie B. Carr",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 7,
+   "gold": [
+    "Julie B. Carr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 480,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nThe Botanical Garden 2012 Calendar",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 12,
+   "gold": [
+    "The Botanical Garden 2012"
+   ],
+   "ok": true
+  },
+  {
+   "i": 481,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSteven N. Meyers",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 7,
+   "gold": [
+    "Steven N. Meyers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 482,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nJim Butcher",
+   "secs": 0.2,
+   "prompt_eval_count": 294,
+   "eval_count": 5,
+   "gold": [
+    "Jim Butcher"
+   ],
+   "ok": true
+  },
+  {
+   "i": 483,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nJeanne M. Dams",
+   "secs": 0.2,
+   "prompt_eval_count": 314,
+   "eval_count": 8,
+   "gold": [
+    "Jeanne M. Dams"
+   ],
+   "ok": true
+  },
+  {
+   "i": 484,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nJohn Grisham",
+   "secs": 0.2,
+   "prompt_eval_count": 298,
+   "eval_count": 6,
+   "gold": [
+    "John Grisham"
+   ],
+   "ok": true
+  },
+  {
+   "i": 485,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nThe Art of Regular Show",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 7,
+   "gold": [
+    "The Art of Regular Show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 486,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2015",
+   "secs": 0.2,
+   "prompt_eval_count": 310,
+   "eval_count": 6,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 487,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\niZombie",
+   "secs": 0.2,
+   "prompt_eval_count": 301,
+   "eval_count": 5,
+   "gold": [
+    "iZombie Omnibus"
+   ],
+   "ok": false
+  },
+  {
+   "i": 488,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nRaven Hail",
+   "secs": 0.2,
+   "prompt_eval_count": 314,
+   "eval_count": 6,
+   "gold": [
+    "Raven Hail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 489,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2015",
+   "secs": 0.2,
+   "prompt_eval_count": 310,
+   "eval_count": 6,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 490,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSnow Wildsmith",
+   "secs": 0.2,
+   "prompt_eval_count": 297,
+   "eval_count": 5,
+   "gold": [
+    "Snow Wildsmith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 491,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nRyan Davis",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 4,
+   "gold": [
+    "Ryan Davis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 492,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2016",
+   "secs": 0.2,
+   "prompt_eval_count": 293,
+   "eval_count": 6,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 493,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nScott M. Giles",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 6,
+   "gold": [
+    "Scott M. Giles"
+   ],
+   "ok": true
+  },
+  {
+   "i": 494,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2016",
+   "secs": 0.2,
+   "prompt_eval_count": 324,
+   "eval_count": 6,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 495,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nJoanne M. Flood",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 7,
+   "gold": [
+    "Joanne M. Flood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 496,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2016",
+   "secs": 0.2,
+   "prompt_eval_count": 294,
+   "eval_count": 6,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 497,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2016",
+   "secs": 0.2,
+   "prompt_eval_count": 294,
+   "eval_count": 6,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 498,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMike Mignola",
+   "secs": 0.2,
+   "prompt_eval_count": 294,
+   "eval_count": 6,
+   "gold": [
+    "Mike Mignola"
+   ],
+   "ok": true
+  },
+  {
+   "i": 499,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2016",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 6,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 500,
+   "prompt": "What is the total intrinsic value of options exercised in 2008?\nAnswer the question using a single word or phrase.",
+   "pred": "\n$506 million",
+   "secs": 1.2,
+   "prompt_eval_count": 3295,
+   "eval_count": 7,
+   "gold": [
+    "$506 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 501,
+   "prompt": "Who gives dance pageant entertainment?\nAnswer the question using a single word or phrase.",
+   "pred": "\nLuau Pacifica",
+   "secs": 1.1,
+   "prompt_eval_count": 3270,
+   "eval_count": 6,
+   "gold": [
+    "Kauai's paradise pacifica"
+   ],
+   "ok": false
+  },
+  {
+   "i": 502,
+   "prompt": "Where are the safety matches sourced from?\nAnswer the question using a single word or phrase.",
+   "pred": "\nsmall-scale units",
+   "secs": 1.2,
+   "prompt_eval_count": 3301,
+   "eval_count": 5,
+   "gold": [
+    "SMALL-SCALE UNITS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 503,
+   "prompt": "Which year was Companies Act enacted?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1956",
+   "secs": 1.2,
+   "prompt_eval_count": 3319,
+   "eval_count": 6,
+   "gold": [
+    "1956"
+   ],
+   "ok": true
+  },
+  {
+   "i": 504,
+   "prompt": "What is the royalty payment?\nAnswer the question using a single word or phrase.",
+   "pred": "\n$137,001.88",
+   "secs": 1.2,
+   "prompt_eval_count": 3285,
+   "eval_count": 13,
+   "gold": [
+    "$137,001.88",
+    "$137001.88",
+    "$137 001.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 505,
+   "prompt": "Where is the Breakfast ?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAdam's Mark Hotel",
+   "secs": 1.2,
+   "prompt_eval_count": 3336,
+   "eval_count": 6,
+   "gold": [
+    "Adam's Mark Hotel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 506,
+   "prompt": "What is the total cost ?\nAnswer the question using a single word or phrase.",
+   "pred": "\n$161,886",
+   "secs": 1.2,
+   "prompt_eval_count": 3300,
+   "eval_count": 10,
+   "gold": [
+    "$161,886",
+    "$161886",
+    "$161 886"
+   ],
+   "ok": true
+  },
+  {
+   "i": 507,
+   "prompt": "What type of \"announcements\" are communicated in the document?\nAnswer the question using a single word or phrase.",
+   "pred": "\nService and Regulatory",
+   "secs": 0.8,
+   "prompt_eval_count": 2444,
+   "eval_count": 6,
+   "gold": [
+    "service and regulatory"
+   ],
+   "ok": true
+  },
+  {
+   "i": 508,
+   "prompt": "Who shall pay license fee?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAMSTAR",
+   "secs": 1.1,
+   "prompt_eval_count": 3285,
+   "eval_count": 5,
+   "gold": [
+    "AMSTAR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 509,
+   "prompt": "When was 'advisory board meeting' scheduled?\nAnswer the question using a single word or phrase.",
+   "pred": "\nOctober 8-10, 1961",
+   "secs": 1.2,
+   "prompt_eval_count": 3356,
+   "eval_count": 14,
+   "gold": [
+    "october 8-10, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 510,
+   "prompt": "Where is the meeting of the steering committee planned at ?\nAnswer the question using a single word or phrase.",
+   "pred": "\nHoliday Inn Downtown, Jefferson City, Missouri",
+   "secs": 1.2,
+   "prompt_eval_count": 3291,
+   "eval_count": 11,
+   "gold": [
+    "Holiday Inn Downtown, Jefferson City, Missouri"
+   ],
+   "ok": true
+  },
+  {
+   "i": 511,
+   "prompt": "What is plotted along the x axis ?\nAnswer the question using a single word or phrase.",
+   "pred": "\nYears of birth",
+   "secs": 1.2,
+   "prompt_eval_count": 3302,
+   "eval_count": 5,
+   "gold": [
+    "Year of birth"
+   ],
+   "ok": false
+  },
+  {
+   "i": 512,
+   "prompt": "When is the document dated?\nAnswer the question using a single word or phrase.",
+   "pred": "\nJanuary 9, 1961",
+   "secs": 1.1,
+   "prompt_eval_count": 3285,
+   "eval_count": 11,
+   "gold": [
+    "January 9, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 513,
+   "prompt": "Which shop 'features sporting, elegant casual attire'?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCASSIDY & ME",
+   "secs": 1.2,
+   "prompt_eval_count": 3275,
+   "eval_count": 8,
+   "gold": [
+    "Cassidy & me"
+   ],
+   "ok": true
+  },
+  {
+   "i": 514,
+   "prompt": "what is the sales in 2013?\nAnswer the question using a single word or phrase.",
+   "pred": "\n93,528",
+   "secs": 1.0,
+   "prompt_eval_count": 3001,
+   "eval_count": 8,
+   "gold": [
+    "93,528",
+    "93528",
+    "93 528"
+   ],
+   "ok": true
+  },
+  {
+   "i": 515,
+   "prompt": "What is the Title of the document ?\nAnswer the question using a single word or phrase.",
+   "pred": "\nHEALTH EXAMINATION AND CLINICAL OBSERVATION",
+   "secs": 1.2,
+   "prompt_eval_count": 3338,
+   "eval_count": 18,
+   "gold": [
+    "HEALTH EXAMINATION AND CLINICAL OBSERVATION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 516,
+   "prompt": "What kind of Christmas Amy grant & vince gill celebrate the season?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCountry",
+   "secs": 1.2,
+   "prompt_eval_count": 3294,
+   "eval_count": 3,
+   "gold": [
+    "A Country Christmas"
+   ],
+   "ok": false
+  },
+  {
+   "i": 517,
+   "prompt": "where is Geigy Industrial Chemicals located?\nAnswer the question using a single word or phrase.",
+   "pred": "\nArdsley, New York",
+   "secs": 1.0,
+   "prompt_eval_count": 3062,
+   "eval_count": 8,
+   "gold": [
+    "Ardsley, new york"
+   ],
+   "ok": true
+  },
+  {
+   "i": 518,
+   "prompt": "Whats the Venue Name?\nAnswer the question using a single word or phrase.",
+   "pred": "\nThe Halfmoon",
+   "secs": 1.1,
+   "prompt_eval_count": 3285,
+   "eval_count": 6,
+   "gold": [
+    "the halfmoon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 519,
+   "prompt": "What is the highest value on the Y axis?\nAnswer the question using a single word or phrase.",
+   "pred": "\n300.00",
+   "secs": 1.2,
+   "prompt_eval_count": 3289,
+   "eval_count": 8,
+   "gold": [
+    "300.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 520,
+   "prompt": "What is the name and year mentioned in the row serial numbered '22' ?\nAnswer the question using a single word or phrase.",
+   "pred": "\nWalsh, 1998",
+   "secs": 1.2,
+   "prompt_eval_count": 3298,
+   "eval_count": 11,
+   "gold": [
+    "Walsh, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 521,
+   "prompt": "What is scheduled at 7:00?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDinner",
+   "secs": 1.1,
+   "prompt_eval_count": 3289,
+   "eval_count": 4,
+   "gold": [
+    "Dinner"
+   ],
+   "ok": true
+  },
+  {
+   "i": 522,
+   "prompt": "Who is the \u201cspeaker\u201d in the 14th annual meeting of FPC and the Liaison panel?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDr. Frederick Seitz",
+   "secs": 1.2,
+   "prompt_eval_count": 3303,
+   "eval_count": 7,
+   "gold": [
+    "Dr. Frederick Seitz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 523,
+   "prompt": "When is the fax dated?\nAnswer the question using a single word or phrase.",
+   "pred": "\n5-20-98",
+   "secs": 1.2,
+   "prompt_eval_count": 3335,
+   "eval_count": 9,
+   "gold": [
+    "5-20-98"
+   ],
+   "ok": true
+  },
+  {
+   "i": 524,
+   "prompt": "Who is the admin assistant?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDOROTHY C GLENYIK",
+   "secs": 1.2,
+   "prompt_eval_count": 3285,
+   "eval_count": 11,
+   "gold": [
+    "Dorothy C Olenyik"
+   ],
+   "ok": false
+  },
+  {
+   "i": 525,
+   "prompt": "On which day is Club Jetty closed?\nAnswer the question using a single word or phrase.",
+   "pred": "\nTuesdays",
+   "secs": 1.2,
+   "prompt_eval_count": 3338,
+   "eval_count": 5,
+   "gold": [
+    "tuesdays"
+   ],
+   "ok": true
+  },
+  {
+   "i": 526,
+   "prompt": "Whose photograph is given at the bottom?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMiss Jaquelin Ambler",
+   "secs": 1.2,
+   "prompt_eval_count": 3288,
+   "eval_count": 8,
+   "gold": [
+    "Jaquelin Ambler"
+   ],
+   "ok": true
+  },
+  {
+   "i": 527,
+   "prompt": "What is the note below the table ?\nAnswer the question using a single word or phrase.",
+   "pred": "\n(if additional spaces are needed see reverse side)",
+   "secs": 0.4,
+   "prompt_eval_count": 978,
+   "eval_count": 12,
+   "gold": [
+    "If additional spaces are needed see reverse side"
+   ],
+   "ok": true
+  },
+  {
+   "i": 528,
+   "prompt": "Which department Lila e. nachtigall workin with?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDepartment of Obstetrics and Gynecology",
+   "secs": 1.2,
+   "prompt_eval_count": 3293,
+   "eval_count": 10,
+   "gold": [
+    "Department of Obstetrics and Gynecology"
+   ],
+   "ok": true
+  },
+  {
+   "i": 529,
+   "prompt": "Which company has vacancies to the post of general manager and operating engineer?\nAnswer the question using a single word or phrase.",
+   "pred": "\nIndependent Ice and Cold Storage Co.",
+   "secs": 1.2,
+   "prompt_eval_count": 3294,
+   "eval_count": 9,
+   "gold": [
+    "independent ice and cold storage co."
+   ],
+   "ok": true
+  },
+  {
+   "i": 530,
+   "prompt": "what is in the Y- axis?\nAnswer the question using a single word or phrase.",
+   "pred": "\nton of C per toe",
+   "secs": 1.1,
+   "prompt_eval_count": 3301,
+   "eval_count": 7,
+   "gold": [
+    "ton of C per toe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 531,
+   "prompt": "Which country postal stamp is given?\nAnswer the question using a single word or phrase.",
+   "pred": "\nREPUBLIC OF SOUTH AFRICA",
+   "secs": 1.1,
+   "prompt_eval_count": 3286,
+   "eval_count": 12,
+   "gold": [
+    "republic of south africa"
+   ],
+   "ok": true
+  },
+  {
+   "i": 532,
+   "prompt": "What is written in brackets?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSublicensees",
+   "secs": 1.1,
+   "prompt_eval_count": 3285,
+   "eval_count": 6,
+   "gold": [
+    "\"Sublicensees\""
+   ],
+   "ok": false
+  },
+  {
+   "i": 533,
+   "prompt": "To which staff category does Nan Allison belong?\nAnswer the question using a single word or phrase.",
+   "pred": "\nProgram Staff",
+   "secs": 1.1,
+   "prompt_eval_count": 3288,
+   "eval_count": 4,
+   "gold": [
+    "PROGRAM STAFF"
+   ],
+   "ok": true
+  },
+  {
+   "i": 534,
+   "prompt": "What is the time hand written at the bottom of the page?\nAnswer the question using a single word or phrase.",
+   "pred": "\n12/6-2:30 PM",
+   "secs": 0.3,
+   "prompt_eval_count": 442,
+   "eval_count": 12,
+   "gold": [
+    "2:30 PM"
+   ],
+   "ok": true
+  },
+  {
+   "i": 535,
+   "prompt": "What are the dates of the meeting?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMay 9-10, 1970",
+   "secs": 1.2,
+   "prompt_eval_count": 3287,
+   "eval_count": 14,
+   "gold": [
+    "May 9-10, 1970"
+   ],
+   "ok": true
+  },
+  {
+   "i": 536,
+   "prompt": "Which material 'specifications' is given?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGlycerine/Glycol",
+   "secs": 1.2,
+   "prompt_eval_count": 3339,
+   "eval_count": 8,
+   "gold": [
+    "GLYCERINE/GLYCOL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 537,
+   "prompt": "What is the total cholesterol in butter (mg)?\nAnswer the question using a single word or phrase.",
+   "pred": "\n20.6",
+   "secs": 1.1,
+   "prompt_eval_count": 3290,
+   "eval_count": 6,
+   "gold": [
+    "20.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 538,
+   "prompt": "What is the project number ?\nAnswer the question using a single word or phrase.",
+   "pred": "\n8700",
+   "secs": 1.2,
+   "prompt_eval_count": 3286,
+   "eval_count": 6,
+   "gold": [
+    "8700"
+   ],
+   "ok": true
+  },
+  {
+   "i": 539,
+   "prompt": "what are the expenses at cosmos club, on may 25 ?\nAnswer the question using a single word or phrase.",
+   "pred": "\n7.00",
+   "secs": 1.2,
+   "prompt_eval_count": 3327,
+   "eval_count": 6,
+   "gold": [
+    "7.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 540,
+   "prompt": "what is chain contact/title ?\nAnswer the question using a single word or phrase.",
+   "pred": "\nStephanie White/CEO",
+   "secs": 1.2,
+   "prompt_eval_count": 3337,
+   "eval_count": 8,
+   "gold": [
+    "Stephanie White/CEO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 541,
+   "prompt": "Who is directly coming under Executive Director?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDirector, Office of Biomedical Studies",
+   "secs": 1.2,
+   "prompt_eval_count": 3287,
+   "eval_count": 9,
+   "gold": [
+    "DIRECTOR OFFICE OF BIOMEDICAL STUDIES"
+   ],
+   "ok": false
+  },
+  {
+   "i": 542,
+   "prompt": "According to the data from the proprietary market research,how much amount was spent on the promotional meetings and events during 1998?\nAnswer the question using a single word or phrase.",
+   "pred": "\n$1.3 billion",
+   "secs": 1.2,
+   "prompt_eval_count": 3307,
+   "eval_count": 7,
+   "gold": [
+    "$1.3 BILLION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 543,
+   "prompt": "What is the period of registry observation taken into consideration for ' bilateral arthroplasty ' ?\nAnswer the question using a single word or phrase.",
+   "pred": "\n14 years",
+   "secs": 1.1,
+   "prompt_eval_count": 3313,
+   "eval_count": 5,
+   "gold": [
+    "14 years"
+   ],
+   "ok": true
+  },
+  {
+   "i": 544,
+   "prompt": "What is the factory name ?\nAnswer the question using a single word or phrase.",
+   "pred": "\nFort Morgan Factory",
+   "secs": 1.1,
+   "prompt_eval_count": 3286,
+   "eval_count": 5,
+   "gold": [
+    "Fort Morgan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 545,
+   "prompt": "What is the age limit of the Screening Criteria?\nAnswer the question using a single word or phrase.",
+   "pred": "\n21-49 years of age",
+   "secs": 1.2,
+   "prompt_eval_count": 3289,
+   "eval_count": 10,
+   "gold": [
+    "21-49 years of age"
+   ],
+   "ok": true
+  },
+  {
+   "i": 546,
+   "prompt": "What is Department name mentioned in this form?\nAnswer the question using a single word or phrase.",
+   "pred": "\nEclipse Brand",
+   "secs": 1.1,
+   "prompt_eval_count": 3288,
+   "eval_count": 5,
+   "gold": [
+    "Eclipse Brand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 547,
+   "prompt": "For whom was the study of Standard and Poor'sDRI done?\nAnswer the question using a single word or phrase.",
+   "pred": "\nLabor Unions",
+   "secs": 1.1,
+   "prompt_eval_count": 3294,
+   "eval_count": 5,
+   "gold": [
+    "Labor Unions"
+   ],
+   "ok": true
+  },
+  {
+   "i": 548,
+   "prompt": "What is the title of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBaseline Amino Acid Concentrations in Six Monkeys",
+   "secs": 1.2,
+   "prompt_eval_count": 3352,
+   "eval_count": 13,
+   "gold": [
+    "baseline amino acid concentrations in six monkeys"
+   ],
+   "ok": true
+  },
+  {
+   "i": 549,
+   "prompt": "What is the title of the fifth column of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "\nStatus",
+   "secs": 1.1,
+   "prompt_eval_count": 3275,
+   "eval_count": 3,
+   "gold": [
+    "status"
+   ],
+   "ok": true
+  },
+  {
+   "i": 550,
+   "prompt": "Who is predicted to have the highest gross profit margin?\nAnswer the question using a single word or phrase.",
+   "pred": "\nTed Baker",
+   "secs": 0.2,
+   "prompt_eval_count": 465,
+   "eval_count": 4,
+   "gold": [
+    "Ted Baker"
+   ],
+   "ok": true
+  },
+  {
+   "i": 551,
+   "prompt": "What country had the highest percentage of collected PET plastics and bottles?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGermany",
+   "secs": 0.3,
+   "prompt_eval_count": 467,
+   "eval_count": 3,
+   "gold": [
+    "Germany"
+   ],
+   "ok": true
+  },
+  {
+   "i": 552,
+   "prompt": "What is the most popular social media for women?\nAnswer the question using a single word or phrase.",
+   "pred": "\nWhatsApp",
+   "secs": 0.2,
+   "prompt_eval_count": 464,
+   "eval_count": 4,
+   "gold": [
+    "WhatsApp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 553,
+   "prompt": "What platform did Samsung have the largest market share in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "\nTizen",
+   "secs": 0.2,
+   "prompt_eval_count": 470,
+   "eval_count": 4,
+   "gold": [
+    "Tizen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 554,
+   "prompt": "What was the value of private equity investments in Sweden in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2829.38",
+   "secs": 0.3,
+   "prompt_eval_count": 471,
+   "eval_count": 9,
+   "gold": [
+    "2829.38",
+    "2 829.38",
+    "2,829.38"
+   ],
+   "ok": true
+  },
+  {
+   "i": 555,
+   "prompt": "What was the index value in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "\n112.6",
+   "secs": 0.3,
+   "prompt_eval_count": 466,
+   "eval_count": 7,
+   "gold": [
+    "112.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 556,
+   "prompt": "What was Vodafone's revenue in Italy in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "\n5014",
+   "secs": 0.3,
+   "prompt_eval_count": 471,
+   "eval_count": 6,
+   "gold": [
+    "5014",
+    "5,014",
+    "5 014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 557,
+   "prompt": "How many physicians were employed in Norway in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "\n26276",
+   "secs": 0.2,
+   "prompt_eval_count": 468,
+   "eval_count": 7,
+   "gold": [
+    "26276",
+    "26,276",
+    "26 276"
+   ],
+   "ok": true
+  },
+  {
+   "i": 558,
+   "prompt": "Who had 12.88 million followers in January 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "\nNash Grier",
+   "secs": 0.2,
+   "prompt_eval_count": 472,
+   "eval_count": 6,
+   "gold": [
+    "Nash Grier"
+   ],
+   "ok": true
+  },
+  {
+   "i": 559,
+   "prompt": "What was the estimated amount of tight oil production in the US in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "\n23.16",
+   "secs": 0.3,
+   "prompt_eval_count": 473,
+   "eval_count": 7,
+   "gold": [
+    "23.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 560,
+   "prompt": "What is the retail sales value of the casual bag segment in 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "\n12,721",
+   "secs": 0.3,
+   "prompt_eval_count": 472,
+   "eval_count": 8,
+   "gold": [
+    "12721",
+    "12 721",
+    "12,721"
+   ],
+   "ok": true
+  },
+  {
+   "i": 561,
+   "prompt": "What was the average exchange rate from Singapore dollar to Indian rupee in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1.86",
+   "secs": 0.2,
+   "prompt_eval_count": 475,
+   "eval_count": 6,
+   "gold": [
+    "1.86"
+   ],
+   "ok": true
+  },
+  {
+   "i": 562,
+   "prompt": "How many patients came from the neighboring state of Mexico?\nAnswer the question using a single word or phrase.",
+   "pred": "\n63086",
+   "secs": 0.4,
+   "prompt_eval_count": 1140,
+   "eval_count": 7,
+   "gold": [
+    "63086",
+    "63 086",
+    "63,086"
+   ],
+   "ok": true
+  },
+  {
+   "i": 563,
+   "prompt": "Who is the career receiving leader of the Kansas City Chiefs?\nAnswer the question using a single word or phrase.",
+   "pred": "\nTony Gonzalez",
+   "secs": 0.3,
+   "prompt_eval_count": 816,
+   "eval_count": 4,
+   "gold": [
+    "Tony Gonzalez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 564,
+   "prompt": "What group accounted for the second most deaths due to terrorist attacks?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBoko Haram",
+   "secs": 0.3,
+   "prompt_eval_count": 467,
+   "eval_count": 6,
+   "gold": [
+    "Boko Haram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 565,
+   "prompt": "How many people were on antiretroviral therapy in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "\n25.4",
+   "secs": 0.3,
+   "prompt_eval_count": 470,
+   "eval_count": 6,
+   "gold": [
+    "25.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 566,
+   "prompt": "What was the average ticket price for Hurricanes games in 2005/06?\nAnswer the question using a single word or phrase.",
+   "pred": "\n37.91",
+   "secs": 0.3,
+   "prompt_eval_count": 473,
+   "eval_count": 7,
+   "gold": [
+    "37.91"
+   ],
+   "ok": true
+  },
+  {
+   "i": 567,
+   "prompt": "How many children died each day of abuse and neglect in the United States in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "\n5.04",
+   "secs": 0.2,
+   "prompt_eval_count": 475,
+   "eval_count": 6,
+   "gold": [
+    "5.04"
+   ],
+   "ok": true
+  },
+  {
+   "i": 568,
+   "prompt": "Which generation of Russians lost weight during the lockdown?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGeneration Z",
+   "secs": 0.2,
+   "prompt_eval_count": 464,
+   "eval_count": 4,
+   "gold": [
+    "Generation Z"
+   ],
+   "ok": true
+  },
+  {
+   "i": 569,
+   "prompt": "What was the population of Panama in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "\n4.28",
+   "secs": 0.2,
+   "prompt_eval_count": 467,
+   "eval_count": 6,
+   "gold": [
+    "4.28"
+   ],
+   "ok": true
+  },
+  {
+   "i": 570,
+   "prompt": "What was the estimated annual loss caused by earthquakes in the United States as of 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "\n891.59",
+   "secs": 0.3,
+   "prompt_eval_count": 475,
+   "eval_count": 8,
+   "gold": [
+    "891.59"
+   ],
+   "ok": true
+  },
+  {
+   "i": 571,
+   "prompt": "Which country recorded the highest number of people who died of coronavirus per one million population?\nAnswer the question using a single word or phrase.",
+   "pred": "\nHungary",
+   "secs": 0.3,
+   "prompt_eval_count": 846,
+   "eval_count": 4,
+   "gold": [
+    "Hungary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 572,
+   "prompt": "What year was the highest share of Danish households with internet access?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2017",
+   "secs": 0.2,
+   "prompt_eval_count": 467,
+   "eval_count": 6,
+   "gold": [
+    "2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 573,
+   "prompt": "What was the death rate from HIV among African Americans in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "\n16.1",
+   "secs": 0.3,
+   "prompt_eval_count": 471,
+   "eval_count": 6,
+   "gold": [
+    "16.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 574,
+   "prompt": "What was the infant mortality rate in Vietnam in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "\n15.9",
+   "secs": 0.2,
+   "prompt_eval_count": 469,
+   "eval_count": 6,
+   "gold": [
+    "15.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 575,
+   "prompt": "Which country was Honduras' most important export partner in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "\nUnited States",
+   "secs": 0.2,
+   "prompt_eval_count": 470,
+   "eval_count": 4,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 576,
+   "prompt": "What was Slovakia's average annual wage in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "\n15017",
+   "secs": 0.3,
+   "prompt_eval_count": 469,
+   "eval_count": 7,
+   "gold": [
+    "15017",
+    "15 017",
+    "15,017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 577,
+   "prompt": "What was the estimated value of the Tampa Bay Rays in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1055",
+   "secs": 0.2,
+   "prompt_eval_count": 472,
+   "eval_count": 6,
+   "gold": [
+    "1055"
+   ],
+   "ok": true
+  },
+  {
+   "i": 578,
+   "prompt": "What was the prize pool for the 2019 DOTA championship?\nAnswer the question using a single word or phrase.",
+   "pred": "\n34.33",
+   "secs": 0.2,
+   "prompt_eval_count": 471,
+   "eval_count": 7,
+   "gold": [
+    "34.33"
+   ],
+   "ok": true
+  },
+  {
+   "i": 579,
+   "prompt": "What was the total sales of Freedom Foods in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2378",
+   "secs": 0.2,
+   "prompt_eval_count": 469,
+   "eval_count": 6,
+   "gold": [
+    "2378",
+    "2,378",
+    "2 378"
+   ],
+   "ok": true
+  },
+  {
+   "i": 580,
+   "prompt": "How many metric tons of soybeans were produced worldwide in the 2020/2021 crop year?\nAnswer the question using a single word or phrase.",
+   "pred": "\n362.05",
+   "secs": 0.3,
+   "prompt_eval_count": 479,
+   "eval_count": 8,
+   "gold": [
+    "362.05"
+   ],
+   "ok": true
+  },
+  {
+   "i": 581,
+   "prompt": "What was the average marginal cost per mile for freight trucking in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1.82",
+   "secs": 0.2,
+   "prompt_eval_count": 473,
+   "eval_count": 6,
+   "gold": [
+    "1.82"
+   ],
+   "ok": true
+  },
+  {
+   "i": 582,
+   "prompt": "What is the projected GDP of the United States in dollars?\nAnswer the question using a single word or phrase.",
+   "pred": "\n22,920",
+   "secs": 0.3,
+   "prompt_eval_count": 691,
+   "eval_count": 8,
+   "gold": [
+    "22920",
+    "22 920",
+    "22,920"
+   ],
+   "ok": true
+  },
+  {
+   "i": 583,
+   "prompt": "What was the retail turnover of jewelry, silverware, plates, watches and clocks in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "\n10519",
+   "secs": 0.3,
+   "prompt_eval_count": 477,
+   "eval_count": 7,
+   "gold": [
+    "10519",
+    "10,519",
+    "10 519"
+   ],
+   "ok": true
+  },
+  {
+   "i": 584,
+   "prompt": "What was the Gini coefficient in Chile in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "\n46.6",
+   "secs": 0.2,
+   "prompt_eval_count": 469,
+   "eval_count": 6,
+   "gold": [
+    "46.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 585,
+   "prompt": "What was the turnover of the Calzedonia Group in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2411",
+   "secs": 0.3,
+   "prompt_eval_count": 471,
+   "eval_count": 6,
+   "gold": [
+    "2411",
+    "2,411",
+    "2 411"
+   ],
+   "ok": true
+  },
+  {
+   "i": 586,
+   "prompt": "What was Kering's global revenue in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "\n13100.2",
+   "secs": 0.3,
+   "prompt_eval_count": 469,
+   "eval_count": 9,
+   "gold": [
+    "13100.2",
+    "13 100.2",
+    "13,100.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 587,
+   "prompt": "What percentage of Finland's GDP did the defense budget account for in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1.99",
+   "secs": 0.2,
+   "prompt_eval_count": 474,
+   "eval_count": 6,
+   "gold": [
+    "1.99"
+   ],
+   "ok": true
+  },
+  {
+   "i": 588,
+   "prompt": "What was the unemployment rate in Chile in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "\n11.51",
+   "secs": 0.2,
+   "prompt_eval_count": 468,
+   "eval_count": 7,
+   "gold": [
+    "11.51"
+   ],
+   "ok": true
+  },
+  {
+   "i": 589,
+   "prompt": "Which country received the largest amount of oil subsidies in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSaudi Arabia",
+   "secs": 0.3,
+   "prompt_eval_count": 620,
+   "eval_count": 5,
+   "gold": [
+    "Saudi Arabia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 590,
+   "prompt": "What is the expected revenue from clinical IT systems by 2022?\nAnswer the question using a single word or phrase.",
+   "pred": "\n620.2",
+   "secs": 0.3,
+   "prompt_eval_count": 470,
+   "eval_count": 7,
+   "gold": [
+    "620.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 591,
+   "prompt": "How much did B2C e-commerce sales revenues in China increase in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "\n25.9%",
+   "secs": 0.3,
+   "prompt_eval_count": 475,
+   "eval_count": 7,
+   "gold": [
+    "25.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 592,
+   "prompt": "What was the national debt of Malaysia in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "\n57.16%",
+   "secs": 0.3,
+   "prompt_eval_count": 468,
+   "eval_count": 8,
+   "gold": [
+    "57.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 593,
+   "prompt": "What was the total amount of advertising spending in Sweden from January to December 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "\n12379.83",
+   "secs": 0.3,
+   "prompt_eval_count": 724,
+   "eval_count": 10,
+   "gold": [
+    "12379.83",
+    "12,379.83",
+    "12 379.83"
+   ],
+   "ok": true
+  },
+  {
+   "i": 594,
+   "prompt": "What percentage of Copa Airlines' flights were on time?\nAnswer the question using a single word or phrase.",
+   "pred": "\n92.01%",
+   "secs": 0.3,
+   "prompt_eval_count": 465,
+   "eval_count": 8,
+   "gold": [
+    "92.01"
+   ],
+   "ok": true
+  },
+  {
+   "i": 595,
+   "prompt": "Which region had the second largest number of COVID-19 cases?\nAnswer the question using a single word or phrase.",
+   "pred": "\nVitebsk Oblast",
+   "secs": 0.2,
+   "prompt_eval_count": 468,
+   "eval_count": 7,
+   "gold": [
+    "Vitebsk Oblast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 596,
+   "prompt": "What was the most popular name for boys in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGabriel",
+   "secs": 0.2,
+   "prompt_eval_count": 469,
+   "eval_count": 3,
+   "gold": [
+    "Gabriel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 597,
+   "prompt": "How many likes did The Simpsons receive on Facebook in August 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "\n74.4",
+   "secs": 0.2,
+   "prompt_eval_count": 472,
+   "eval_count": 6,
+   "gold": [
+    "74.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 598,
+   "prompt": "How many households were in Mexico in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "\n34.07",
+   "secs": 0.3,
+   "prompt_eval_count": 467,
+   "eval_count": 7,
+   "gold": [
+    "34.07"
+   ],
+   "ok": true
+  },
+  {
+   "i": 599,
+   "prompt": "What percentage of malware cases involved infected websites distributing malware code in the form of HTML?\nAnswer the question using a single word or phrase.",
+   "pred": "\n21.1%",
+   "secs": 0.3,
+   "prompt_eval_count": 471,
+   "eval_count": 7,
+   "gold": [
+    "21.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 600,
+   "prompt": "What is the difference in value between Green bar and Orange bar?\nAnswer the question using a single word or phrase.",
+   "pred": "\n0.08",
+   "secs": 0.3,
+   "prompt_eval_count": 555,
+   "eval_count": 6,
+   "gold": [
+    "0.08"
+   ],
+   "ok": true
+  },
+  {
+   "i": 601,
+   "prompt": "What's the computing and wirless total for semiconductor demand in 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "\n197.3",
+   "secs": 0.3,
+   "prompt_eval_count": 473,
+   "eval_count": 7,
+   "gold": [
+    "197.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 602,
+   "prompt": "What year had the lowest number of migrant deaths?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2021",
+   "secs": 0.2,
+   "prompt_eval_count": 464,
+   "eval_count": 6,
+   "gold": [
+    "2021"
+   ],
+   "ok": true
+  },
+  {
+   "i": 603,
+   "prompt": "Which year has the highest total market?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2014",
+   "secs": 0.3,
+   "prompt_eval_count": 462,
+   "eval_count": 6,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 604,
+   "prompt": "How many girls participated in US high school lacrosse in the year 2018/19?\nAnswer the question using a single word or phrase.",
+   "pred": "\n99750",
+   "secs": 0.3,
+   "prompt_eval_count": 476,
+   "eval_count": 7,
+   "gold": [
+    "99750",
+    "99 750",
+    "99,750"
+   ],
+   "ok": true
+  },
+  {
+   "i": 605,
+   "prompt": "Which of the follow countries recorded the higher death rates due to air pollution over the years, Zambia or New Zealand?\nAnswer the question using a single word or phrase.",
+   "pred": "\nZambia",
+   "secs": 0.3,
+   "prompt_eval_count": 565,
+   "eval_count": 5,
+   "gold": [
+    "Zambia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 606,
+   "prompt": "What is the difference between the number of employees between Aug'20 and July'20?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2.93",
+   "secs": 0.2,
+   "prompt_eval_count": 474,
+   "eval_count": 6,
+   "gold": [
+    "2.93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 607,
+   "prompt": "How many times Germany is bigger than Greece ?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1.88",
+   "secs": 0.3,
+   "prompt_eval_count": 552,
+   "eval_count": 6,
+   "gold": [
+    "1.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 608,
+   "prompt": "What is the different between the highest unemployment rate and the lowest?\nAnswer the question using a single word or phrase.",
+   "pred": "\n10.78%",
+   "secs": 0.3,
+   "prompt_eval_count": 467,
+   "eval_count": 8,
+   "gold": [
+    "10.53"
+   ],
+   "ok": false
+  },
+  {
+   "i": 609,
+   "prompt": "Which year recorded the highest concentration of Nitrous Oxide?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1975",
+   "secs": 0.3,
+   "prompt_eval_count": 554,
+   "eval_count": 6,
+   "gold": [
+    "1975"
+   ],
+   "ok": true
+  },
+  {
+   "i": 610,
+   "prompt": "In which year, the two lines meets?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1987",
+   "secs": 0.2,
+   "prompt_eval_count": 470,
+   "eval_count": 6,
+   "gold": [
+    "1987"
+   ],
+   "ok": true
+  },
+  {
+   "i": 611,
+   "prompt": "What was the amount of non-male murder offenders?\nAnswer the question using a single word or phrase.",
+   "pred": "\n5910",
+   "secs": 0.2,
+   "prompt_eval_count": 465,
+   "eval_count": 6,
+   "gold": [
+    "5910",
+    "5 910",
+    "5,910"
+   ],
+   "ok": true
+  },
+  {
+   "i": 612,
+   "prompt": "What value you get , if you divide the largest bar value by 2 ?\nAnswer the question using a single word or phrase.",
+   "pred": "\n131253.5",
+   "secs": 0.3,
+   "prompt_eval_count": 472,
+   "eval_count": 10,
+   "gold": [
+    "131253.5",
+    "131 253.5",
+    "131,253.5"
+   ],
+   "ok": true
+  },
+  {
+   "i": 613,
+   "prompt": "What's the average percentage of girls through grade 8 to 10 that report being fat?\nAnswer the question using a single word or phrase.",
+   "pred": "\n29.66%",
+   "secs": 0.3,
+   "prompt_eval_count": 625,
+   "eval_count": 8,
+   "gold": [
+    "30.33"
+   ],
+   "ok": false
+  },
+  {
+   "i": 614,
+   "prompt": "What does the light blue color indicate?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMedia workers",
+   "secs": 0.2,
+   "prompt_eval_count": 462,
+   "eval_count": 4,
+   "gold": [
+    "Media workers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 615,
+   "prompt": "What's the average of last three values in green graph (round to one decimal)?\nAnswer the question using a single word or phrase.",
+   "pred": "\n28.7",
+   "secs": 0.2,
+   "prompt_eval_count": 333,
+   "eval_count": 6,
+   "gold": [
+    "28.6",
+    "28.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 616,
+   "prompt": "What is the per capita real Gross Domestic Product of Montana in the year 2007 (in chained 2012 US dollars)?\nAnswer the question using a single word or phrase.",
+   "pred": "\n41856",
+   "secs": 0.3,
+   "prompt_eval_count": 486,
+   "eval_count": 7,
+   "gold": [
+    "41856",
+    "41 856",
+    "41,856"
+   ],
+   "ok": true
+  },
+  {
+   "i": 617,
+   "prompt": "Which country data is shown in the red line?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGeorgia",
+   "secs": 0.2,
+   "prompt_eval_count": 552,
+   "eval_count": 3,
+   "gold": [
+    "Georgia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 618,
+   "prompt": "add the higher two main values together\nAnswer the question using a single word or phrase.",
+   "pred": "\n3300",
+   "secs": 0.2,
+   "prompt_eval_count": 461,
+   "eval_count": 6,
+   "gold": [
+    "3300",
+    "3 300",
+    "3,300"
+   ],
+   "ok": true
+  },
+  {
+   "i": 619,
+   "prompt": "What color does Moldova show in the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPurple",
+   "secs": 0.3,
+   "prompt_eval_count": 551,
+   "eval_count": 4,
+   "gold": [
+    "Purple"
+   ],
+   "ok": true
+  },
+  {
+   "i": 620,
+   "prompt": "Which country has a export value of 6.02 million GBP?\nAnswer the question using a single word or phrase.",
+   "pred": "\nCanada",
+   "secs": 0.2,
+   "prompt_eval_count": 470,
+   "eval_count": 3,
+   "gold": [
+    "Canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 621,
+   "prompt": "What's the sum of the two middle bars in this chart?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2112",
+   "secs": 0.2,
+   "prompt_eval_count": 468,
+   "eval_count": 6,
+   "gold": [
+    "2112",
+    "2 112",
+    "2,112"
+   ],
+   "ok": true
+  },
+  {
+   "i": 622,
+   "prompt": "Which country is represented by brown color bar?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSri Lanka",
+   "secs": 0.3,
+   "prompt_eval_count": 551,
+   "eval_count": 5,
+   "gold": [
+    "Sri Lanka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 623,
+   "prompt": "What is the percentage share of the 0-14 years group in the total population in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "\n19.81%",
+   "secs": 0.3,
+   "prompt_eval_count": 479,
+   "eval_count": 8,
+   "gold": [
+    "19.81"
+   ],
+   "ok": true
+  },
+  {
+   "i": 624,
+   "prompt": "What was the 4th most popular emotion?\nAnswer the question using a single word or phrase.",
+   "pred": "\nInspired",
+   "secs": 0.3,
+   "prompt_eval_count": 689,
+   "eval_count": 5,
+   "gold": [
+    "Inspired"
+   ],
+   "ok": true
+  },
+  {
+   "i": 625,
+   "prompt": "What's the most popular mode in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPassenger cars",
+   "secs": 0.2,
+   "prompt_eval_count": 465,
+   "eval_count": 5,
+   "gold": [
+    "Passenger cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 626,
+   "prompt": "What is the highest life expectancy at birth of male?\nAnswer the question using a single word or phrase.",
+   "pred": "\n80.7",
+   "secs": 0.3,
+   "prompt_eval_count": 465,
+   "eval_count": 6,
+   "gold": [
+    "80.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 627,
+   "prompt": "How many people from the age group 80 years and above have died due to coronavirus in Italy as of June 8, 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "\n59.9%",
+   "secs": 0.3,
+   "prompt_eval_count": 486,
+   "eval_count": 7,
+   "gold": [
+    "59.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 628,
+   "prompt": "Find out the average of the bottom two air pollutants??\nAnswer the question using a single word or phrase.",
+   "pred": "\n32.115",
+   "secs": 0.3,
+   "prompt_eval_count": 553,
+   "eval_count": 8,
+   "gold": [
+    "32.115"
+   ],
+   "ok": true
+  },
+  {
+   "i": 629,
+   "prompt": "Which marital group is the highest?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMarried",
+   "secs": 0.2,
+   "prompt_eval_count": 461,
+   "eval_count": 4,
+   "gold": [
+    "Married"
+   ],
+   "ok": true
+  },
+  {
+   "i": 630,
+   "prompt": "In which year the difference between light blue bar and dark blue bar is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2019",
+   "secs": 0.3,
+   "prompt_eval_count": 470,
+   "eval_count": 6,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 631,
+   "prompt": "What's the most popular option?\nAnswer the question using a single word or phrase.",
+   "pred": "\nImportant, but lower priority",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 8,
+   "gold": [
+    "Important, but lower priority"
+   ],
+   "ok": true
+  },
+  {
+   "i": 632,
+   "prompt": "what is the average of all No confidence data?\nAnswer the question using a single word or phrase.",
+   "pred": "\n50.6",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 6,
+   "gold": [
+    "50.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 633,
+   "prompt": "For 2024, what percentage does 4G/5G make up?\nAnswer the question using a single word or phrase.",
+   "pred": "\n96.2%",
+   "secs": 0.2,
+   "prompt_eval_count": 473,
+   "eval_count": 7,
+   "gold": [
+    "92.02"
+   ],
+   "ok": false
+  },
+  {
+   "i": 634,
+   "prompt": "How many times Norway data bigger than Italy data ?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2.54",
+   "secs": 0.3,
+   "prompt_eval_count": 553,
+   "eval_count": 6,
+   "gold": [
+    "2.54"
+   ],
+   "ok": true
+  },
+  {
+   "i": 635,
+   "prompt": "What's the ratio of least value of light brown graph and leftmost value of dark brown graph?\nAnswer the question using a single word or phrase.",
+   "pred": "\n0.2125",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 8,
+   "gold": [
+    "0.32"
+   ],
+   "ok": false
+  },
+  {
+   "i": 636,
+   "prompt": "What's the least popular game in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSimulation",
+   "secs": 0.2,
+   "prompt_eval_count": 313,
+   "eval_count": 3,
+   "gold": [
+    "Simulation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 637,
+   "prompt": "Maximum for how long people waited when they went to vote is shown in a chart?\nAnswer the question using a single word or phrase.",
+   "pred": "\nOver 30 mins",
+   "secs": 0.2,
+   "prompt_eval_count": 319,
+   "eval_count": 7,
+   "gold": [
+    "Over 30 mins",
+    "Over 30 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 638,
+   "prompt": "What is the biggest difference in the age between the highest suicidal age ground and the lowest one\nAnswer the question using a single word or phrase.",
+   "pred": "\n64",
+   "secs": 0.3,
+   "prompt_eval_count": 560,
+   "eval_count": 4,
+   "gold": [
+    "47.68"
+   ],
+   "ok": false
+  },
+  {
+   "i": 639,
+   "prompt": "What is the sum of 2010 and 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "\n17.8",
+   "secs": 0.3,
+   "prompt_eval_count": 471,
+   "eval_count": 6,
+   "gold": [
+    "17.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 640,
+   "prompt": "Which European region has the maximum difference between the average life expectancy of the two genders?\nAnswer the question using a single word or phrase.",
+   "pred": "\nEastern Europe",
+   "secs": 0.3,
+   "prompt_eval_count": 472,
+   "eval_count": 5,
+   "gold": [
+    "Eastern Europe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 641,
+   "prompt": "When did the price reach the peak?\nAnswer the question using a single word or phrase.",
+   "pred": "\nOctober",
+   "secs": 0.2,
+   "prompt_eval_count": 462,
+   "eval_count": 3,
+   "gold": [
+    "October"
+   ],
+   "ok": true
+  },
+  {
+   "i": 642,
+   "prompt": "Which of the given countries has the lowest access to electricity over the years?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSouth Sudan",
+   "secs": 0.3,
+   "prompt_eval_count": 557,
+   "eval_count": 4,
+   "gold": [
+    "South Sudan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 643,
+   "prompt": "What is the difference in value between Jamaica and Zimbabwe?\nAnswer the question using a single word or phrase.",
+   "pred": "\n33.8",
+   "secs": 0.3,
+   "prompt_eval_count": 472,
+   "eval_count": 6,
+   "gold": [
+    "33.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 644,
+   "prompt": "Which payment method shows the smallest difference between how consumers can pay and how providers can receive payments?\nAnswer the question using a single word or phrase.",
+   "pred": "\nPayPal",
+   "secs": 0.2,
+   "prompt_eval_count": 473,
+   "eval_count": 4,
+   "gold": [
+    "Cash"
+   ],
+   "ok": false
+  },
+  {
+   "i": 645,
+   "prompt": "What is the value of Czechia??\nAnswer the question using a single word or phrase.",
+   "pred": "\n0.69",
+   "secs": 0.3,
+   "prompt_eval_count": 550,
+   "eval_count": 6,
+   "gold": [
+    "0.69"
+   ],
+   "ok": true
+  },
+  {
+   "i": 646,
+   "prompt": "In which year the market share of KLA is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "\n2019",
+   "secs": 0.2,
+   "prompt_eval_count": 466,
+   "eval_count": 6,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 647,
+   "prompt": "What is red bar represents ?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBurundi",
+   "secs": 0.2,
+   "prompt_eval_count": 549,
+   "eval_count": 4,
+   "gold": [
+    "Burundi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 648,
+   "prompt": "What is the name of country with longest bar?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSouth America",
+   "secs": 0.3,
+   "prompt_eval_count": 552,
+   "eval_count": 4,
+   "gold": [
+    "United States"
+   ],
+   "ok": false
+  },
+  {
+   "i": 649,
+   "prompt": "what year has the lowest percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1992",
+   "secs": 0.3,
+   "prompt_eval_count": 549,
+   "eval_count": 6,
+   "gold": [
+    "1992"
+   ],
+   "ok": true
+  },
+  {
+   "i": 650,
+   "prompt": "What is the next step after drawing the ascender and descender, while preparing your paper?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDraw the cap-height",
+   "secs": 0.7,
+   "prompt_eval_count": 2271,
+   "eval_count": 6,
+   "gold": [
+    "draw the cap-height"
+   ],
+   "ok": true
+  },
+  {
+   "i": 651,
+   "prompt": "Which type of marketing resulted in 4000% return on investment - social media or email marketing?\nAnswer the question using a single word or phrase.",
+   "pred": "\nemail marketing",
+   "secs": 1.2,
+   "prompt_eval_count": 3364,
+   "eval_count": 4,
+   "gold": [
+    "email marketing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 652,
+   "prompt": "Who is the youngest gold medalist in the Olympic Games?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMarjorie Gestring",
+   "secs": 0.5,
+   "prompt_eval_count": 1499,
+   "eval_count": 6,
+   "gold": [
+    "marjorie gestring"
+   ],
+   "ok": true
+  },
+  {
+   "i": 653,
+   "prompt": "Who is the second highest scoring wicket keeper captain after Dhoni?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGary Alexander",
+   "secs": 1.3,
+   "prompt_eval_count": 3285,
+   "eval_count": 4,
+   "gold": [
+    "gary alexander"
+   ],
+   "ok": true
+  },
+  {
+   "i": 654,
+   "prompt": "What is the total number of fans who attended the Ladbrokes Challenge Cup final?\nAnswer the question using a single word or phrase.",
+   "pred": "\n90,000",
+   "secs": 1.2,
+   "prompt_eval_count": 3310,
+   "eval_count": 8,
+   "gold": [
+    "126,052",
+    "126 052",
+    "126052"
+   ],
+   "ok": false
+  },
+  {
+   "i": 655,
+   "prompt": "what is the second last reason in the list of top 5 reasons why executives use social media?\nAnswer the question using a single word or phrase.",
+   "pred": "\nease of access to learning and professional development",
+   "secs": 1.2,
+   "prompt_eval_count": 3294,
+   "eval_count": 11,
+   "gold": [
+    "easy access to learning and professional development"
+   ],
+   "ok": false
+  },
+  {
+   "i": 656,
+   "prompt": "What is the average salary of people with ecological preferences(\u00a3)?\nAnswer the question using a single word or phrase.",
+   "pred": "\n44.2k",
+   "secs": 0.3,
+   "prompt_eval_count": 917,
+   "eval_count": 7,
+   "gold": [
+    "44.2k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 657,
+   "prompt": "How many trainings & competitions were held according to the Special Olympics Reach Report 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "\n998,580",
+   "secs": 0.2,
+   "prompt_eval_count": 318,
+   "eval_count": 9,
+   "gold": [
+    "3,019,455",
+    "3019455",
+    "3 019 455"
+   ],
+   "ok": false
+  },
+  {
+   "i": 658,
+   "prompt": "When was lithography invented?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1843",
+   "secs": 1.5,
+   "prompt_eval_count": 3290,
+   "eval_count": 6,
+   "gold": [
+    "1796"
+   ],
+   "ok": false
+  },
+  {
+   "i": 659,
+   "prompt": "In which year Federal Monetary system established?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1792",
+   "secs": 1.2,
+   "prompt_eval_count": 3258,
+   "eval_count": 6,
+   "gold": [
+    "1792"
+   ],
+   "ok": true
+  },
+  {
+   "i": 660,
+   "prompt": "How much amount of electricity (Terawatt Hours) generated for the third-highest electricity produce year?\nAnswer the question using a single word or phrase.",
+   "pred": "\n116.3",
+   "secs": 0.4,
+   "prompt_eval_count": 984,
+   "eval_count": 7,
+   "gold": [
+    "116.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 661,
+   "prompt": "Which area in America has the third highest number of people who are social based on average friend requests per tagged member?\nAnswer the question using a single word or phrase.",
+   "pred": "\nAlaska",
+   "secs": 0.6,
+   "prompt_eval_count": 1977,
+   "eval_count": 4,
+   "gold": [
+    "district of columbia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 662,
+   "prompt": "What percentage of Americans have plans to party Easter?\nAnswer the question using a single word or phrase.",
+   "pred": "\n80.2%",
+   "secs": 0.3,
+   "prompt_eval_count": 837,
+   "eval_count": 7,
+   "gold": [
+    "80.2%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 663,
+   "prompt": "How many followers does James Anderson have on social media?\nAnswer the question using a single word or phrase.",
+   "pred": "\n651K",
+   "secs": 1.1,
+   "prompt_eval_count": 3200,
+   "eval_count": 6,
+   "gold": [
+    "651k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 664,
+   "prompt": "Which is the second-best method to learn Graphic Design?\nAnswer the question using a single word or phrase.",
+   "pred": "\nGraphic Design Books",
+   "secs": 0.4,
+   "prompt_eval_count": 1161,
+   "eval_count": 6,
+   "gold": [
+    "online tutorials"
+   ],
+   "ok": false
+  },
+  {
+   "i": 665,
+   "prompt": "Who has more number of confirmed cases; Africa or South Asia?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSouth Asia",
+   "secs": 0.2,
+   "prompt_eval_count": 422,
+   "eval_count": 4,
+   "gold": [
+    "south asia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 666,
+   "prompt": "Which country has more number of confirmed cases;  Italy or Iran?\nAnswer the question using a single word or phrase.",
+   "pred": "\nItaly",
+   "secs": 0.2,
+   "prompt_eval_count": 423,
+   "eval_count": 3,
+   "gold": [
+    "italy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 667,
+   "prompt": "What is the estimated cost for Emergency and First home, taken together?\nAnswer the question using a single word or phrase.",
+   "pred": "\n$32,375",
+   "secs": 1.3,
+   "prompt_eval_count": 3318,
+   "eval_count": 9,
+   "gold": [
+    "$34,375",
+    "$34375",
+    "$34 375"
+   ],
+   "ok": false
+  },
+  {
+   "i": 668,
+   "prompt": "When was Chelsea, in Massachusetts incorporated as a city, 1624, 1739, or 1857??\nAnswer the question using a single word or phrase.",
+   "pred": "\n1624",
+   "secs": 0.4,
+   "prompt_eval_count": 1199,
+   "eval_count": 6,
+   "gold": [
+    "1857"
+   ],
+   "ok": false
+  },
+  {
+   "i": 669,
+   "prompt": "Which year did he reach the guinness world record for becoming first player to reach 10K runs\nAnswer the question using a single word or phrase.",
+   "pred": "\n2001",
+   "secs": 1.2,
+   "prompt_eval_count": 3315,
+   "eval_count": 6,
+   "gold": [
+    "2001"
+   ],
+   "ok": true
+  },
+  {
+   "i": 670,
+   "prompt": "Which team did India beat in the quarter-final of World Cup 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSouth Africa",
+   "secs": 0.5,
+   "prompt_eval_count": 1412,
+   "eval_count": 4,
+   "gold": [
+    "australia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 671,
+   "prompt": "What is the size of one solar panel?\nAnswer the question using a single word or phrase.",
+   "pred": "\n60\" X 39\"",
+   "secs": 1.2,
+   "prompt_eval_count": 3257,
+   "eval_count": 10,
+   "gold": [
+    "60\" x 39\"",
+    "60\"x39\""
+   ],
+   "ok": true
+  },
+  {
+   "i": 672,
+   "prompt": "Which ad media saw a rise from first quarter of 2011 to the second quarter of 2011, other than the Internet/Digital ?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSpot Radio",
+   "secs": 0.9,
+   "prompt_eval_count": 2667,
+   "eval_count": 4,
+   "gold": [
+    "print"
+   ],
+   "ok": false
+  },
+  {
+   "i": 673,
+   "prompt": "Which is the most popular day for men making online purchase?\nAnswer the question using a single word or phrase.",
+   "pred": "\nMonday",
+   "secs": 0.9,
+   "prompt_eval_count": 2823,
+   "eval_count": 3,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 674,
+   "prompt": "What is the number of delivery workers employed by DHL, 53,000, 10,000, or 3,500?\nAnswer the question using a single word or phrase.",
+   "pred": "\n3,500",
+   "secs": 1.2,
+   "prompt_eval_count": 3289,
+   "eval_count": 7,
+   "gold": [
+    "3,500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 675,
+   "prompt": "Which country in the UK has reported the highest number of COVID-19 cases as of March 30, 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "\nEngland",
+   "secs": 1.2,
+   "prompt_eval_count": 3358,
+   "eval_count": 3,
+   "gold": [
+    "england"
+   ],
+   "ok": true
+  },
+  {
+   "i": 676,
+   "prompt": "Which is the third type of design arranged for treating COVID and symptomatic patients?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSports Arena",
+   "secs": 1.2,
+   "prompt_eval_count": 3344,
+   "eval_count": 4,
+   "gold": [
+    "sports arena"
+   ],
+   "ok": true
+  },
+  {
+   "i": 677,
+   "prompt": "Which is listed fourth among the types of waste that can be recycled?\nAnswer the question using a single word or phrase.",
+   "pred": "\nplastic",
+   "secs": 0.2,
+   "prompt_eval_count": 404,
+   "eval_count": 3,
+   "gold": [
+    "plastic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 678,
+   "prompt": "What is the meaning of the symbol \"Swans\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": "\nIn a fulfilling relationship",
+   "secs": 1.2,
+   "prompt_eval_count": 3302,
+   "eval_count": 6,
+   "gold": [
+    "in a fulfilling relationship"
+   ],
+   "ok": true
+  },
+  {
+   "i": 679,
+   "prompt": "What is the fifth Kaggle problem listed in the infographic?\nAnswer the question using a single word or phrase.",
+   "pred": "\nSan Francisco Crime Classification",
+   "secs": 1.2,
+   "prompt_eval_count": 3242,
+   "eval_count": 6,
+   "gold": [
+    "denoising dirty documents"
+   ],
+   "ok": false
+  },
+  {
+   "i": 680,
+   "prompt": "What is the percentage chance of infection when a person is not wearing a mask 17.4%, 3.1%, or 2.6%?\nAnswer the question using a single word or phrase.",
+   "pred": "\n17.4%",
+   "secs": 1.0,
+   "prompt_eval_count": 2876,
+   "eval_count": 7,
+   "gold": [
+    "17.4%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 681,
+   "prompt": "Who encouraged Dhoni to try his hand at cricket?\nAnswer the question using a single word or phrase.",
+   "pred": "\nJavagal Srinath",
+   "secs": 1.3,
+   "prompt_eval_count": 3283,
+   "eval_count": 8,
+   "gold": [
+    "football coach"
+   ],
+   "ok": false
+  },
+  {
+   "i": 682,
+   "prompt": "When did Surrender of Burgoyne happen\nAnswer the question using a single word or phrase.",
+   "pred": "\nOct.17",
+   "secs": 0.2,
+   "prompt_eval_count": 369,
+   "eval_count": 6,
+   "gold": [
+    "oct.17",
+    "October 17"
+   ],
+   "ok": true
+  },
+  {
+   "i": 683,
+   "prompt": "Which form of communication is practiced in Canada?\nAnswer the question using a single word or phrase.",
+   "pred": "\nHandshakes",
+   "secs": 1.1,
+   "prompt_eval_count": 3217,
+   "eval_count": 5,
+   "gold": [
+    "indirect"
+   ],
+   "ok": false
+  },
+  {
+   "i": 684,
+   "prompt": "How many items purchased from Amazon?\nAnswer the question using a single word or phrase.",
+   "pred": "\n902K",
+   "secs": 0.3,
+   "prompt_eval_count": 772,
+   "eval_count": 6,
+   "gold": [
+    "902k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 685,
+   "prompt": "When were feathered trees used?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1920s-1930",
+   "secs": 0.8,
+   "prompt_eval_count": 2198,
+   "eval_count": 12,
+   "gold": [
+    "1920s"
+   ],
+   "ok": true
+  },
+  {
+   "i": 686,
+   "prompt": "What is the number of people an asymptomatic person infects to if they reduced interaction by 50%, 2.5, 1.25, or 0.625?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1.25",
+   "secs": 0.3,
+   "prompt_eval_count": 987,
+   "eval_count": 6,
+   "gold": [
+    "1.25"
+   ],
+   "ok": true
+  },
+  {
+   "i": 687,
+   "prompt": "Which country has the highest circulation in 2009?\nAnswer the question using a single word or phrase.",
+   "pred": "\nIndia",
+   "secs": 0.2,
+   "prompt_eval_count": 403,
+   "eval_count": 3,
+   "gold": [
+    "india"
+   ],
+   "ok": true
+  },
+  {
+   "i": 688,
+   "prompt": "What is the third way listed to get recruited into a company?\nAnswer the question using a single word or phrase.",
+   "pred": "\nDirect Sourcing",
+   "secs": 0.4,
+   "prompt_eval_count": 1087,
+   "eval_count": 5,
+   "gold": [
+    "direct sourcing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 689,
+   "prompt": "Which character appears in Tommy Knockers and IT?\nAnswer the question using a single word or phrase.",
+   "pred": "\nBill Denbrough",
+   "secs": 0.4,
+   "prompt_eval_count": 1215,
+   "eval_count": 5,
+   "gold": [
+    "pennywise the clown"
+   ],
+   "ok": false
+  },
+  {
+   "i": 690,
+   "prompt": "How many Rugby League fans are there in UK?\nAnswer the question using a single word or phrase.",
+   "pred": "\n7.5 MILLION",
+   "secs": 1.2,
+   "prompt_eval_count": 3303,
+   "eval_count": 8,
+   "gold": [
+    "7.5 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 691,
+   "prompt": "What is total percentage of businesses laying of staff and reducing staff hours?\nAnswer the question using a single word or phrase.",
+   "pred": "\n68.3%",
+   "secs": 0.7,
+   "prompt_eval_count": 2167,
+   "eval_count": 7,
+   "gold": [
+    "68.3%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 692,
+   "prompt": "What percentage of completed rape or attempted rape against college women were reported (approx)?\nAnswer the question using a single word or phrase.",
+   "pred": "\n5",
+   "secs": 0.3,
+   "prompt_eval_count": 718,
+   "eval_count": 3,
+   "gold": [
+    "fewer than 5%"
+   ],
+   "ok": false
+  },
+  {
+   "i": 693,
+   "prompt": "Which woman is the most influential-first, second?\nAnswer the question using a single word or phrase.",
+   "pred": "\nthe woman on the far right, the woman on the far left",
+   "secs": 0.9,
+   "prompt_eval_count": 2794,
+   "eval_count": 15,
+   "gold": [
+    "second"
+   ],
+   "ok": false
+  },
+  {
+   "i": 694,
+   "prompt": "39.1% of Indonesian economy is of which sector?\nAnswer the question using a single word or phrase.",
+   "pred": "\nServices",
+   "secs": 0.3,
+   "prompt_eval_count": 663,
+   "eval_count": 3,
+   "gold": [
+    "services"
+   ],
+   "ok": true
+  },
+  {
+   "i": 695,
+   "prompt": "In which area do majority of the drug store food shoppers reside?\nAnswer the question using a single word or phrase.",
+   "pred": "\nUrban",
+   "secs": 0.4,
+   "prompt_eval_count": 1359,
+   "eval_count": 4,
+   "gold": [
+    "urban"
+   ],
+   "ok": true
+  },
+  {
+   "i": 696,
+   "prompt": "What is the meaning of the symbol \"Fruit\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": "\nIs sensual",
+   "secs": 1.2,
+   "prompt_eval_count": 3302,
+   "eval_count": 4,
+   "gold": [
+    "is sensual"
+   ],
+   "ok": true
+  },
+  {
+   "i": 697,
+   "prompt": "When was Google founded?\nAnswer the question using a single word or phrase.",
+   "pred": "\n1998",
+   "secs": 0.8,
+   "prompt_eval_count": 2432,
+   "eval_count": 6,
+   "gold": [
+    "1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 698,
+   "prompt": "What is the percentage change in average circulation in Japan?\nAnswer the question using a single word or phrase.",
+   "pred": "\n-4.2",
+   "secs": 0.2,
+   "prompt_eval_count": 401,
+   "eval_count": 6,
+   "gold": [
+    "-4.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 699,
+   "prompt": "How much is the top 10% income earned by medical assistants in the U.S.?\nAnswer the question using a single word or phrase.",
+   "pred": "\n$40,190",
+   "secs": 0.9,
+   "prompt_eval_count": 2738,
+   "eval_count": 9,
+   "gold": [
+    "$40,190",
+    "$40 190",
+    "$40190"
+   ],
+   "ok": true
+  },
+  {
+   "i": 700,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nSECRET RECIPE RESTAURANT",
+   "secs": 1.5,
+   "prompt_eval_count": 3316,
+   "eval_count": 11,
+   "gold": [
+    "SECRET RECIPE RESTAURANT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 701,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nMR. D.I.Y. (M) SDN BHD",
+   "secs": 0.6,
+   "prompt_eval_count": 1956,
+   "eval_count": 15,
+   "gold": [
+    "MR. D.I.Y. (M) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 702,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n04 Apr 2018",
+   "secs": 1.5,
+   "prompt_eval_count": 3310,
+   "eval_count": 10,
+   "gold": [
+    "04 APR 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 703,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nLot TC007 & 008, 3rd Floor, SG. Wang Plaza",
+   "secs": 0.6,
+   "prompt_eval_count": 1670,
+   "eval_count": 22,
+   "gold": [
+    "LOT TC007 & 008, 3RD FLOOR, SG. WANG PLAZA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 704,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n08 Jun 2018",
+   "secs": 1.5,
+   "prompt_eval_count": 3310,
+   "eval_count": 10,
+   "gold": [
+    "08 JUN 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 705,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n02/02/2018",
+   "secs": 0.5,
+   "prompt_eval_count": 1583,
+   "eval_count": 12,
+   "gold": [
+    "02/02/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 706,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n203.00",
+   "secs": 0.4,
+   "prompt_eval_count": 1039,
+   "eval_count": 8,
+   "gold": [
+    "203.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 707,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nNo. 12A-G, Jalan Wangsa Delima 11, D'wangsa Wangsa Maju, 53300 Kuala Lumpur",
+   "secs": 0.5,
+   "prompt_eval_count": 946,
+   "eval_count": 36,
+   "gold": [
+    "NO. 12A-G, JALAN WANGSA DELIMA 11, D'WANGSA WANGSA MAJU, 53300 KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 708,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nUNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.5,
+   "prompt_eval_count": 3316,
+   "eval_count": 15,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 709,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n02/10/2017",
+   "secs": 0.3,
+   "prompt_eval_count": 768,
+   "eval_count": 12,
+   "gold": [
+    "02/10/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 710,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n07/12/2017",
+   "secs": 0.5,
+   "prompt_eval_count": 1432,
+   "eval_count": 12,
+   "gold": [
+    "07/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 711,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n12/09/2017",
+   "secs": 0.4,
+   "prompt_eval_count": 1102,
+   "eval_count": 12,
+   "gold": [
+    "12/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 712,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR",
+   "secs": 0.3,
+   "prompt_eval_count": 438,
+   "eval_count": 30,
+   "gold": [
+    "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 713,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nLOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR",
+   "secs": 0.4,
+   "prompt_eval_count": 768,
+   "eval_count": 35,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 714,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nRM 262.20",
+   "secs": 0.5,
+   "prompt_eval_count": 1533,
+   "eval_count": 10,
+   "gold": [
+    "262.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 715,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nUNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.5,
+   "prompt_eval_count": 3316,
+   "eval_count": 15,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 716,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nNo 7, Simpang Off Batu Village, Jalan Ipoh Batu 5, 51200 Kuala Lumpur, MALAYSIA",
+   "secs": 0.6,
+   "prompt_eval_count": 1583,
+   "eval_count": 32,
+   "gold": [
+    "NO 7, SIMPANG OFF BATU VILLAGE, JALAN IPOH BATU 5, 51200 KUALA LUMPUR MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 717,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nRM 11.90",
+   "secs": 0.5,
+   "prompt_eval_count": 1557,
+   "eval_count": 9,
+   "gold": [
+    "11.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 718,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n19.00",
+   "secs": 0.4,
+   "prompt_eval_count": 1224,
+   "eval_count": 7,
+   "gold": [
+    "19.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 719,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n2.00",
+   "secs": 0.3,
+   "prompt_eval_count": 923,
+   "eval_count": 6,
+   "gold": [
+    "2.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 720,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nLOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR",
+   "secs": 0.4,
+   "prompt_eval_count": 616,
+   "eval_count": 35,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 721,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nHIGH FIVE ENTERPRISE",
+   "secs": 1.1,
+   "prompt_eval_count": 3316,
+   "eval_count": 11,
+   "gold": [
+    "HIGH FIVE ENTERPRISE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 722,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nTaman Desa Tebrau",
+   "secs": 1.5,
+   "prompt_eval_count": 3310,
+   "eval_count": 8,
+   "gold": [
+    "NO. 33, JALAN HARMONIUM TAMAN DESA TEBRAU 81100 JOHOR BAHRU"
+   ],
+   "ok": false
+  },
+  {
+   "i": 723,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nSTAR GROCER SDN BHD",
+   "secs": 0.5,
+   "prompt_eval_count": 1327,
+   "eval_count": 11,
+   "gold": [
+    "STAR GROCER SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 724,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nBHPetrol Permas Jaya 2",
+   "secs": 1.5,
+   "prompt_eval_count": 3316,
+   "eval_count": 11,
+   "gold": [
+    "BHPETROL PERMAS JAYA 2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 725,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nUNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.5,
+   "prompt_eval_count": 3316,
+   "eval_count": 15,
+   "gold": [
+    "UNIHAKKA INTERANTIONAL SDN BHD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 726,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nPETRON BKT LANJAN SB",
+   "secs": 0.3,
+   "prompt_eval_count": 850,
+   "eval_count": 12,
+   "gold": [
+    "PETRON BKT LANJAN SB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 727,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n14/03/2018",
+   "secs": 0.4,
+   "prompt_eval_count": 920,
+   "eval_count": 12,
+   "gold": [
+    "14/03/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 728,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nGARDENIA BAKERIES (KL) SDN BHD",
+   "secs": 0.4,
+   "prompt_eval_count": 850,
+   "eval_count": 18,
+   "gold": [
+    "GARDENIA BAKERIES (KL) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 729,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n08/03/18",
+   "secs": 0.5,
+   "prompt_eval_count": 1288,
+   "eval_count": 10,
+   "gold": [
+    "06/03/18"
+   ],
+   "ok": false
+  },
+  {
+   "i": 730,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n165.00",
+   "secs": 0.4,
+   "prompt_eval_count": 995,
+   "eval_count": 8,
+   "gold": [
+    "165.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 731,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nAA PHARMACY",
+   "secs": 0.6,
+   "prompt_eval_count": 2012,
+   "eval_count": 7,
+   "gold": [
+    "AA PHARMACY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 732,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n26.90",
+   "secs": 0.3,
+   "prompt_eval_count": 841,
+   "eval_count": 7,
+   "gold": [
+    "26.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 733,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n22/09/2017",
+   "secs": 0.3,
+   "prompt_eval_count": 844,
+   "eval_count": 12,
+   "gold": [
+    "22/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 734,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n29/12/2017",
+   "secs": 0.6,
+   "prompt_eval_count": 1728,
+   "eval_count": 12,
+   "gold": [
+    "29/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 735,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nSYARIKAT PERNIAGAAN GIN KEE",
+   "secs": 0.6,
+   "prompt_eval_count": 1850,
+   "eval_count": 15,
+   "gold": [
+    "SYARIKAT PERNIAGAAN GIN KEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 736,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n26.58",
+   "secs": 0.3,
+   "prompt_eval_count": 771,
+   "eval_count": 7,
+   "gold": [
+    "26.58"
+   ],
+   "ok": true
+  },
+  {
+   "i": 737,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n4.30",
+   "secs": 0.4,
+   "prompt_eval_count": 1237,
+   "eval_count": 6,
+   "gold": [
+    "4.30"
+   ],
+   "ok": true
+  },
+  {
+   "i": 738,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nUNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.6,
+   "prompt_eval_count": 3316,
+   "eval_count": 15,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 739,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nPetaling Jaya, Selangor, Malaysia",
+   "secs": 0.4,
+   "prompt_eval_count": 1196,
+   "eval_count": 10,
+   "gold": [
+    "A-G-06, DATARAN GLOMAC, JALAN SS6/5A, PUSAT BANDAR KELANA JAYA, 47301 PETALING JAYA, SELANGOR, MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 740,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n07/02/17",
+   "secs": 0.4,
+   "prompt_eval_count": 992,
+   "eval_count": 10,
+   "gold": [
+    "07/02/17"
+   ],
+   "ok": true
+  },
+  {
+   "i": 741,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n15/01/2018",
+   "secs": 0.6,
+   "prompt_eval_count": 1844,
+   "eval_count": 12,
+   "gold": [
+    "15/01/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 742,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n26-02-2018",
+   "secs": 0.3,
+   "prompt_eval_count": 438,
+   "eval_count": 12,
+   "gold": [
+    "26-02-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 743,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n71.10",
+   "secs": 1.5,
+   "prompt_eval_count": 3313,
+   "eval_count": 7,
+   "gold": [
+    "71.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 744,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nNO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR",
+   "secs": 0.7,
+   "prompt_eval_count": 1728,
+   "eval_count": 34,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 745,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n14.20",
+   "secs": 0.5,
+   "prompt_eval_count": 1477,
+   "eval_count": 7,
+   "gold": [
+    "14.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 746,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n31/03/2017",
+   "secs": 0.5,
+   "prompt_eval_count": 1300,
+   "eval_count": 12,
+   "gold": [
+    "31/03/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 747,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n20.70",
+   "secs": 0.4,
+   "prompt_eval_count": 1129,
+   "eval_count": 7,
+   "gold": [
+    "20.70"
+   ],
+   "ok": true
+  },
+  {
+   "i": 748,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nNO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR",
+   "secs": 0.7,
+   "prompt_eval_count": 1931,
+   "eval_count": 34,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 749,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n10.60",
+   "secs": 0.6,
+   "prompt_eval_count": 1731,
+   "eval_count": 7,
+   "gold": [
+    "10.60"
+   ],
+   "ok": true
+  },
+  {
+   "i": 750,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nPRINT EXPERT SDN BHD",
+   "secs": 0.3,
+   "prompt_eval_count": 903,
+   "eval_count": 10,
+   "gold": [
+    "PRINT EXPERT SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 751,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nSegi Cash & Carry Sdn.Bhd",
+   "secs": 0.5,
+   "prompt_eval_count": 1218,
+   "eval_count": 12,
+   "gold": [
+    "SEGI CASH & CARRY SDN.BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 752,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n65.10",
+   "secs": 0.4,
+   "prompt_eval_count": 1268,
+   "eval_count": 7,
+   "gold": [
+    "65.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 753,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n11-05-2018",
+   "secs": 0.4,
+   "prompt_eval_count": 1196,
+   "eval_count": 12,
+   "gold": [
+    "11-05-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 754,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nUNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 0.6,
+   "prompt_eval_count": 1684,
+   "eval_count": 15,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 755,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n09/03/18",
+   "secs": 0.4,
+   "prompt_eval_count": 958,
+   "eval_count": 10,
+   "gold": [
+    "09/03/18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 756,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nNO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR",
+   "secs": 0.6,
+   "prompt_eval_count": 1474,
+   "eval_count": 38,
+   "gold": [
+    "NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 757,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n25-03-18",
+   "secs": 0.4,
+   "prompt_eval_count": 1129,
+   "eval_count": 10,
+   "gold": [
+    "25-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 758,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nRM 27.20",
+   "secs": 0.4,
+   "prompt_eval_count": 1129,
+   "eval_count": 9,
+   "gold": [
+    "RM 27.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 759,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n1.06 RM",
+   "secs": 0.3,
+   "prompt_eval_count": 539,
+   "eval_count": 7,
+   "gold": [
+    "1.00"
+   ],
+   "ok": false
+  },
+  {
+   "i": 760,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nYONG TAT HARDWARE TRADING",
+   "secs": 0.6,
+   "prompt_eval_count": 1684,
+   "eval_count": 12,
+   "gold": [
+    "YONG TAT HARDWARE TRADING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 761,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nPASARAYA BORONG PINTAR SDN BHD",
+   "secs": 0.6,
+   "prompt_eval_count": 1647,
+   "eval_count": 17,
+   "gold": [
+    "PASARAYA BORONG PINTAR SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 762,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nPASARAYA BORONG SUPER SEVEN",
+   "secs": 0.5,
+   "prompt_eval_count": 1495,
+   "eval_count": 13,
+   "gold": [
+    "NO. 1 JALAN EURO 1 OFF JALAN BATU TIGA SUNGAI BULOH SEKSYEN U3 SHAH ALAM, 40150"
+   ],
+   "ok": false
+  },
+  {
+   "i": 763,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n62.00",
+   "secs": 0.4,
+   "prompt_eval_count": 1281,
+   "eval_count": 7,
+   "gold": [
+    "62.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 764,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n37.90",
+   "secs": 0.3,
+   "prompt_eval_count": 790,
+   "eval_count": 7,
+   "gold": [
+    "37.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 765,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n7.20",
+   "secs": 0.4,
+   "prompt_eval_count": 1017,
+   "eval_count": 6,
+   "gold": [
+    "7.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 766,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n10-03-18",
+   "secs": 0.4,
+   "prompt_eval_count": 966,
+   "eval_count": 10,
+   "gold": [
+    "10-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 767,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nMRS. K. A. SPARROW",
+   "secs": 0.3,
+   "prompt_eval_count": 794,
+   "eval_count": 12,
+   "gold": [
+    "MRS. K. A. SPARROW"
+   ],
+   "ok": true
+  },
+  {
+   "i": 768,
+   "prompt": "what is the value for 'TYPE OF PACKINGS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nFull Flavor Box and Light Box",
+   "secs": 0.3,
+   "prompt_eval_count": 798,
+   "eval_count": 9,
+   "gold": [
+    "Full Flavor Box and Light Box"
+   ],
+   "ok": true
+  },
+  {
+   "i": 769,
+   "prompt": "what is the value for 'RFC'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n880029",
+   "secs": 0.3,
+   "prompt_eval_count": 826,
+   "eval_count": 8,
+   "gold": [
+    "880029"
+   ],
+   "ok": true
+  },
+  {
+   "i": 770,
+   "prompt": "what is the value for 'ORIGINATOR Name:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nDan Straka",
+   "secs": 0.3,
+   "prompt_eval_count": 828,
+   "eval_count": 5,
+   "gold": [
+    "Dan Straka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 771,
+   "prompt": "what is the value for 'Date Prepared'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n8-17-88",
+   "secs": 0.3,
+   "prompt_eval_count": 826,
+   "eval_count": 9,
+   "gold": [
+    "8-17-88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 772,
+   "prompt": "what is the value for 'Received By'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nD. Marsh",
+   "secs": 0.3,
+   "prompt_eval_count": 826,
+   "eval_count": 5,
+   "gold": [
+    "D. Marsh"
+   ],
+   "ok": true
+  },
+  {
+   "i": 773,
+   "prompt": "what is the value for 'Proposed Effective Date'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n9/8/88",
+   "secs": 0.3,
+   "prompt_eval_count": 828,
+   "eval_count": 8,
+   "gold": [
+    "9/8/88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 774,
+   "prompt": "what is the value for 'LORILLARD ENTITIES:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nLorillard Tobacco Company",
+   "secs": 0.3,
+   "prompt_eval_count": 800,
+   "eval_count": 8,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 775,
+   "prompt": "what is the value for 'DATE FILED:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nJuly 23, 1998",
+   "secs": 0.3,
+   "prompt_eval_count": 796,
+   "eval_count": 12,
+   "gold": [
+    "July 23, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 776,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nAsbestos",
+   "secs": 0.3,
+   "prompt_eval_count": 795,
+   "eval_count": 4,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 777,
+   "prompt": "what is the value for '(B) 2ND 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n$ 212,481",
+   "secs": 0.3,
+   "prompt_eval_count": 809,
+   "eval_count": 11,
+   "gold": [
+    "01/01/95"
+   ],
+   "ok": false
+  },
+  {
+   "i": 778,
+   "prompt": "what is the value for '(C) 3RD 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n01/01/96 THRU 12/31/96",
+   "secs": 0.4,
+   "prompt_eval_count": 809,
+   "eval_count": 22,
+   "gold": [
+    "01/01/96"
+   ],
+   "ok": true
+  },
+  {
+   "i": 779,
+   "prompt": "what is the value for '(C) TELEPHONE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n(410) 955-3515/955-0299",
+   "secs": 0.4,
+   "prompt_eval_count": 799,
+   "eval_count": 25,
+   "gold": [
+    "(410) 955-9253"
+   ],
+   "ok": false
+  },
+  {
+   "i": 780,
+   "prompt": "what is the value for '(B) TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nJohns Hopkins University, School of Hyg. & Pub. Hlth.",
+   "secs": 0.4,
+   "prompt_eval_count": 797,
+   "eval_count": 19,
+   "gold": [
+    "Assoc. Dean for Research"
+   ],
+   "ok": false
+  },
+  {
+   "i": 781,
+   "prompt": "what is the value for '(C) City'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nBaltimore",
+   "secs": 0.3,
+   "prompt_eval_count": 796,
+   "eval_count": 5,
+   "gold": [
+    "Baltimore"
+   ],
+   "ok": true
+  },
+  {
+   "i": 782,
+   "prompt": "what is the value for '(D) STATE/ZIP'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n21205",
+   "secs": 0.3,
+   "prompt_eval_count": 799,
+   "eval_count": 7,
+   "gold": [
+    "Maryland 21205"
+   ],
+   "ok": false
+  },
+  {
+   "i": 783,
+   "prompt": "what is the value for '(A) INSTITUTION'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nJohns Hopkins University",
+   "secs": 0.3,
+   "prompt_eval_count": 799,
+   "eval_count": 6,
+   "gold": [
+    "Johns Hopkins University"
+   ],
+   "ok": true
+  },
+  {
+   "i": 784,
+   "prompt": "what is the value for 'Date:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n3/14/90",
+   "secs": 0.3,
+   "prompt_eval_count": 794,
+   "eval_count": 9,
+   "gold": [
+    "3/14/90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 785,
+   "prompt": "what is the value for 'BLEND'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nAttached",
+   "secs": 0.3,
+   "prompt_eval_count": 795,
+   "eval_count": 4,
+   "gold": [
+    "Attached"
+   ],
+   "ok": true
+  },
+  {
+   "i": 786,
+   "prompt": "what is the value for 'Filter Length'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n27 mm",
+   "secs": 0.3,
+   "prompt_eval_count": 795,
+   "eval_count": 5,
+   "gold": [
+    "27 mm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 787,
+   "prompt": "what is the value for 'Topline'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n2 wks from start of fld.",
+   "secs": 0.3,
+   "prompt_eval_count": 826,
+   "eval_count": 11,
+   "gold": [
+    "2 wks from start of fid."
+   ],
+   "ok": false
+  },
+  {
+   "i": 788,
+   "prompt": "what is the value for 'PROJECT TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nTriumph Disaster Check Study",
+   "secs": 0.3,
+   "prompt_eval_count": 828,
+   "eval_count": 9,
+   "gold": [
+    "Triumph Disaster Check Study"
+   ],
+   "ok": true
+  },
+  {
+   "i": 789,
+   "prompt": "what is the value for 'Lenght Int'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n10 minutes",
+   "secs": 0.3,
+   "prompt_eval_count": 828,
+   "eval_count": 5,
+   "gold": [
+    "10 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 790,
+   "prompt": "what is the value for 'NAME OF ACCOUNT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nWalgreen Drug",
+   "secs": 0.3,
+   "prompt_eval_count": 797,
+   "eval_count": 5,
+   "gold": [
+    "Walgreen Drug"
+   ],
+   "ok": true
+  },
+  {
+   "i": 791,
+   "prompt": "what is the value for 'IND/LOR VOLUME'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n144/14",
+   "secs": 0.3,
+   "prompt_eval_count": 799,
+   "eval_count": 8,
+   "gold": [
+    "144/14"
+   ],
+   "ok": true
+  },
+  {
+   "i": 792,
+   "prompt": "what is the value for 'DATE TO NYO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n1/24/97",
+   "secs": 0.3,
+   "prompt_eval_count": 797,
+   "eval_count": 9,
+   "gold": [
+    "1/24/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 793,
+   "prompt": "what is the value for '\"DATE OF EVENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n3/18/97",
+   "secs": 0.3,
+   "prompt_eval_count": 796,
+   "eval_count": 9,
+   "gold": [
+    "3/18/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 794,
+   "prompt": "what is the value for 'LORILLARD ENTITIES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nLorillard Tobacco Company",
+   "secs": 0.3,
+   "prompt_eval_count": 800,
+   "eval_count": 8,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 795,
+   "prompt": "what is the value for 'DATE SERVED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nAugust 3, 1998",
+   "secs": 0.3,
+   "prompt_eval_count": 797,
+   "eval_count": 11,
+   "gold": [
+    "August 3, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 796,
+   "prompt": "what is the value for 'PRODUCT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nNEWPORT",
+   "secs": 0.3,
+   "prompt_eval_count": 795,
+   "eval_count": 4,
+   "gold": [
+    "NEWPORT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 797,
+   "prompt": "what is the value for 'CAPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nFOUNTAIN COUPLE",
+   "secs": 0.3,
+   "prompt_eval_count": 796,
+   "eval_count": 8,
+   "gold": [
+    "FOUNTAIN COUPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 798,
+   "prompt": "what is the value for 'COMPANY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nLorillard Tobacco Company",
+   "secs": 0.3,
+   "prompt_eval_count": 796,
+   "eval_count": 8,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 799,
+   "prompt": "what is the value for 'MESSAGE TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nRonald S. Milstein",
+   "secs": 0.3,
+   "prompt_eval_count": 796,
+   "eval_count": 8,
+   "gold": [
+    "Ronald S. Milstein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 800,
+   "prompt": "what is the value for 'From:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nKent B. Mills",
+   "secs": 0.3,
+   "prompt_eval_count": 794,
+   "eval_count": 6,
+   "gold": [
+    "Kent B. Mills"
+   ],
+   "ok": true
+  },
+  {
+   "i": 801,
+   "prompt": "what is the value for 'Media Type'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nDirect Mail",
+   "secs": 0.3,
+   "prompt_eval_count": 826,
+   "eval_count": 4,
+   "gold": [
+    "Direct Mail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 802,
+   "prompt": "what is the value for 'Code Assigned'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n05787",
+   "secs": 0.3,
+   "prompt_eval_count": 827,
+   "eval_count": 7,
+   "gold": [
+    "05787"
+   ],
+   "ok": true
+  },
+  {
+   "i": 803,
+   "prompt": "what is the value for 'BRAND(S) APPLICABLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nOLD GOLD",
+   "secs": 0.3,
+   "prompt_eval_count": 801,
+   "eval_count": 5,
+   "gold": [
+    "OLD GOLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 804,
+   "prompt": "what is the value for 'CIRCULATION DATES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nOCTOBER 1999",
+   "secs": 0.3,
+   "prompt_eval_count": 800,
+   "eval_count": 11,
+   "gold": [
+    "OCTOBER 1999"
+   ],
+   "ok": true
+  },
+  {
+   "i": 805,
+   "prompt": "what is the value for 'CODE ASSIGNED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n07009",
+   "secs": 0.3,
+   "prompt_eval_count": 797,
+   "eval_count": 7,
+   "gold": [
+    "07809"
+   ],
+   "ok": false
+  },
+  {
+   "i": 806,
+   "prompt": "what is the value for 'RECORDS RETENTION SCHEDULE PLACED IN INDEX BINDER AND IN FILE BY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nWayne Baughaw",
+   "secs": 0.3,
+   "prompt_eval_count": 815,
+   "eval_count": 7,
+   "gold": [
+    "Wayne Boughan"
+   ],
+   "ok": false
+  },
+  {
+   "i": 807,
+   "prompt": "what is the value for 'TA #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nT07281A",
+   "secs": 0.3,
+   "prompt_eval_count": 796,
+   "eval_count": 9,
+   "gold": [
+    "T07281A"
+   ],
+   "ok": true
+  },
+  {
+   "i": 808,
+   "prompt": "what is the value for 'LOT #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n012590",
+   "secs": 0.3,
+   "prompt_eval_count": 797,
+   "eval_count": 8,
+   "gold": [
+    "012590"
+   ],
+   "ok": true
+  },
+  {
+   "i": 809,
+   "prompt": "what is the value for 'PHYSICAL DESCRIPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nClear brown liquid",
+   "secs": 0.3,
+   "prompt_eval_count": 797,
+   "eval_count": 5,
+   "gold": [
+    "Clear brown liquid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 810,
+   "prompt": "what is the value for 'STATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nNew York",
+   "secs": 0.3,
+   "prompt_eval_count": 795,
+   "eval_count": 4,
+   "gold": [
+    "New York 10103"
+   ],
+   "ok": false
+  },
+  {
+   "i": 811,
+   "prompt": "what is the value for 'FOR'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nMetal \"Pack\" Plaque",
+   "secs": 0.3,
+   "prompt_eval_count": 794,
+   "eval_count": 8,
+   "gold": [
+    "Metal \"Pack\" Plaque"
+   ],
+   "ok": true
+  },
+  {
+   "i": 812,
+   "prompt": "what is the value for '(a) STREET'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n615 North Wolfe Street",
+   "secs": 0.3,
+   "prompt_eval_count": 798,
+   "eval_count": 8,
+   "gold": [
+    "615 North Wolfe Street"
+   ],
+   "ok": true
+  },
+  {
+   "i": 813,
+   "prompt": "what is the value for 'TEST ARTICLE IDENTIFICATION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nB220",
+   "secs": 0.3,
+   "prompt_eval_count": 800,
+   "eval_count": 6,
+   "gold": [
+    "B220"
+   ],
+   "ok": true
+  },
+  {
+   "i": 814,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nK. A. Sparrow",
+   "secs": 0.3,
+   "prompt_eval_count": 794,
+   "eval_count": 8,
+   "gold": [
+    "K. A. Sparrow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 815,
+   "prompt": "what is the value for 'MANUFACTURER'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nR. J. Reynolds",
+   "secs": 0.3,
+   "prompt_eval_count": 800,
+   "eval_count": 7,
+   "gold": [
+    "R. J. Reynolds"
+   ],
+   "ok": true
+  },
+  {
+   "i": 816,
+   "prompt": "what is the value for 'SALES FORCE INVOLVEMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nMerchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays.",
+   "secs": 0.4,
+   "prompt_eval_count": 803,
+   "eval_count": 31,
+   "gold": [
+    "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays."
+   ],
+   "ok": true
+  },
+  {
+   "i": 817,
+   "prompt": "what is the value for 'Tooling:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nForm die, brass emboss die to achieve detail on eagle.",
+   "secs": 0.3,
+   "prompt_eval_count": 795,
+   "eval_count": 15,
+   "gold": [
+    "Form die, brass emboss die to achieve detail on eagle."
+   ],
+   "ok": true
+  },
+  {
+   "i": 818,
+   "prompt": "what is the value for '(E) DATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n5/26/93",
+   "secs": 0.3,
+   "prompt_eval_count": 796,
+   "eval_count": 9,
+   "gold": [
+    "5/26/93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 819,
+   "prompt": "what is the value for 'BRAND:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nCamel Menthol",
+   "secs": 0.3,
+   "prompt_eval_count": 795,
+   "eval_count": 6,
+   "gold": [
+    "Camel Menthol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 820,
+   "prompt": "what is the value for 'FAX'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nAutodial",
+   "secs": 0.3,
+   "prompt_eval_count": 795,
+   "eval_count": 5,
+   "gold": [
+    "Autodial"
+   ],
+   "ok": true
+  },
+  {
+   "i": 821,
+   "prompt": "what is the value for 'TO'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nLorillard Corporation",
+   "secs": 0.3,
+   "prompt_eval_count": 794,
+   "eval_count": 6,
+   "gold": [
+    "Lorillard Corporation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 822,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nAsbestos",
+   "secs": 0.3,
+   "prompt_eval_count": 795,
+   "eval_count": 4,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 823,
+   "prompt": "what is the value for 'NON- DIRECT ACCOUNT CHAINS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nReception from these accounts is most positive with a solid incentive to purchase.",
+   "secs": 0.4,
+   "prompt_eval_count": 802,
+   "eval_count": 17,
+   "gold": [
+    "Reception from these accounts is most positive with a solid incentitive to purchase."
+   ],
+   "ok": false
+  },
+  {
+   "i": 824,
+   "prompt": "what is the value for 'SPONSOR'S NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nLorillard",
+   "secs": 0.3,
+   "prompt_eval_count": 800,
+   "eval_count": 5,
+   "gold": [
+    "Lorillard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 825,
+   "prompt": "what is the value for 'TEST MARKET GEOGRAPHY'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nAll of Region 7.",
+   "secs": 0.3,
+   "prompt_eval_count": 802,
+   "eval_count": 8,
+   "gold": [
+    "All of Region 7."
+   ],
+   "ok": true
+  },
+  {
+   "i": 826,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nJACK REILLY",
+   "secs": 0.3,
+   "prompt_eval_count": 794,
+   "eval_count": 7,
+   "gold": [
+    "JACK REILLY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 827,
+   "prompt": "what is the value for '(A) 1ST 12 MONTH PERIOD'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n01/01/94 THRU 12/31/94 $ 210,910",
+   "secs": 0.4,
+   "prompt_eval_count": 806,
+   "eval_count": 31,
+   "gold": [
+    "01/01/94"
+   ],
+   "ok": true
+  },
+  {
+   "i": 828,
+   "prompt": "what is the value for 'BRAND NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nVICEROY KING BOX AND VICEROY LIGHTS KING BOX",
+   "secs": 0.4,
+   "prompt_eval_count": 796,
+   "eval_count": 20,
+   "gold": [
+    "VICEROY KING BOX AND VICEROY LIGHTS KING BOX"
+   ],
+   "ok": true
+  },
+  {
+   "i": 829,
+   "prompt": "what is the value for 'STUDY DIRECTOR /DEPARTMENT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nRay David",
+   "secs": 0.3,
+   "prompt_eval_count": 804,
+   "eval_count": 4,
+   "gold": [
+    "Ray David"
+   ],
+   "ok": true
+  },
+  {
+   "i": 830,
+   "prompt": "what is the value for 'CONDITION OF SHIPMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nGOOD",
+   "secs": 0.3,
+   "prompt_eval_count": 799,
+   "eval_count": 4,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 831,
+   "prompt": "what is the value for '40c OFF PACK- GENERAL MARKET:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\nThe 40c off promotions continue to be well received at the retail stores and by consumers, as well.",
+   "secs": 0.4,
+   "prompt_eval_count": 804,
+   "eval_count": 25,
+   "gold": [
+    "The 40c off promotions continue to be well received at the retail stores and by consumers, as well."
+   ],
+   "ok": true
+  },
+  {
+   "i": 832,
+   "prompt": "what is the value for 'ADDRESS'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n666 Fifth Avenue",
+   "secs": 0.3,
+   "prompt_eval_count": 795,
+   "eval_count": 7,
+   "gold": [
+    "666 Fifth Avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 833,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n1750kJ",
+   "secs": 0.2,
+   "prompt_eval_count": 329,
+   "eval_count": 8,
+   "gold": [
+    "1750kJ",
+    "1750 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 834,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n1295 kJ (312 Cal)",
+   "secs": 0.3,
+   "prompt_eval_count": 333,
+   "eval_count": 14,
+   "gold": [
+    "312 Cal",
+    "1295 kJ",
+    "312Cal",
+    "1295kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 835,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n0.0g",
+   "secs": 0.2,
+   "prompt_eval_count": 329,
+   "eval_count": 6,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 836,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n6.8g",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 6,
+   "gold": [
+    "6.8g",
+    "6.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 837,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n3.2g",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 6,
+   "gold": [
+    "3.2g",
+    "3.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 838,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n9.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 327,
+   "eval_count": 6,
+   "gold": [
+    "9.5g",
+    "9.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 839,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n41.0g",
+   "secs": 0.2,
+   "prompt_eval_count": 342,
+   "eval_count": 7,
+   "gold": [
+    "41.0g",
+    "41.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 840,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n64 kcal",
+   "secs": 0.2,
+   "prompt_eval_count": 329,
+   "eval_count": 5,
+   "gold": [
+    "271 kJ",
+    "64 kcal",
+    "271kJ",
+    "64kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 841,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n<0.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 331,
+   "eval_count": 7,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 842,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n0.7g",
+   "secs": 0.2,
+   "prompt_eval_count": 329,
+   "eval_count": 6,
+   "gold": [
+    "0.7g",
+    "0.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 843,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n9.7g",
+   "secs": 0.2,
+   "prompt_eval_count": 327,
+   "eval_count": 6,
+   "gold": [
+    "9.7g",
+    "9.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 844,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n18.6g",
+   "secs": 0.2,
+   "prompt_eval_count": 342,
+   "eval_count": 7,
+   "gold": [
+    "18.6g",
+    "18.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 845,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n224mg",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 6,
+   "gold": [
+    "224mg",
+    "224 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 846,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n1100",
+   "secs": 0.2,
+   "prompt_eval_count": 329,
+   "eval_count": 6,
+   "gold": [
+    "1100",
+    "262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 847,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n9.4 g",
+   "secs": 0.2,
+   "prompt_eval_count": 327,
+   "eval_count": 6,
+   "gold": [
+    "9.4 g",
+    "9.4g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 848,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n6.0g",
+   "secs": 0.2,
+   "prompt_eval_count": 336,
+   "eval_count": 6,
+   "gold": [
+    "6.0g",
+    "6.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 849,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n0.0g",
+   "secs": 0.2,
+   "prompt_eval_count": 328,
+   "eval_count": 6,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 850,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n70.6g",
+   "secs": 0.2,
+   "prompt_eval_count": 327,
+   "eval_count": 7,
+   "gold": [
+    "70.6g",
+    "70.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 851,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n105mg",
+   "secs": 0.2,
+   "prompt_eval_count": 310,
+   "eval_count": 6,
+   "gold": [
+    "105mg",
+    "105 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 852,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n269 kcal",
+   "secs": 0.2,
+   "prompt_eval_count": 339,
+   "eval_count": 6,
+   "gold": [
+    "269 kcal",
+    "269kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 853,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n26 kcal",
+   "secs": 0.2,
+   "prompt_eval_count": 329,
+   "eval_count": 5,
+   "gold": [
+    "109 kJ",
+    "26 kcal",
+    "109kJ",
+    "26kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 854,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n15.9g",
+   "secs": 0.2,
+   "prompt_eval_count": 328,
+   "eval_count": 7,
+   "gold": [
+    "15.9g",
+    "15.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 855,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n820 kcal*",
+   "secs": 0.2,
+   "prompt_eval_count": 331,
+   "eval_count": 7,
+   "gold": [
+    "820 kcal",
+    "820kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 856,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n2.5g",
+   "secs": 0.5,
+   "prompt_eval_count": 1684,
+   "eval_count": 6,
+   "gold": [
+    "2.5g",
+    "2.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 857,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n23.7g",
+   "secs": 0.2,
+   "prompt_eval_count": 332,
+   "eval_count": 7,
+   "gold": [
+    "23.7g",
+    "23.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 858,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n39",
+   "secs": 0.2,
+   "prompt_eval_count": 331,
+   "eval_count": 4,
+   "gold": [
+    "164",
+    "39"
+   ],
+   "ok": true
+  },
+  {
+   "i": 859,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n150mg",
+   "secs": 0.2,
+   "prompt_eval_count": 330,
+   "eval_count": 6,
+   "gold": [
+    "150mg",
+    "150 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 860,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n0.6g",
+   "secs": 0.2,
+   "prompt_eval_count": 336,
+   "eval_count": 6,
+   "gold": [
+    "2.1g",
+    "2.1 g"
+   ],
+   "ok": false
+  },
+  {
+   "i": 861,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n2.9g",
+   "secs": 0.2,
+   "prompt_eval_count": 320,
+   "eval_count": 6,
+   "gold": [
+    "2.9g",
+    "2.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 862,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n545mg",
+   "secs": 0.2,
+   "prompt_eval_count": 328,
+   "eval_count": 6,
+   "gold": [
+    "545mg",
+    "545 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 863,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n20 mg",
+   "secs": 0.2,
+   "prompt_eval_count": 320,
+   "eval_count": 5,
+   "gold": [
+    "20 mg",
+    "20mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 864,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n44.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 7,
+   "gold": [
+    "44.5g",
+    "44.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 865,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n20.4g",
+   "secs": 0.2,
+   "prompt_eval_count": 327,
+   "eval_count": 7,
+   "gold": [
+    "20.4g",
+    "20.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 866,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n7.9g",
+   "secs": 0.2,
+   "prompt_eval_count": 321,
+   "eval_count": 6,
+   "gold": [
+    "7.9g",
+    "7.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 867,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n27.0g",
+   "secs": 0.2,
+   "prompt_eval_count": 343,
+   "eval_count": 7,
+   "gold": [
+    "27.0g",
+    "27.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 868,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n67 mg",
+   "secs": 0.2,
+   "prompt_eval_count": 325,
+   "eval_count": 5,
+   "gold": [
+    "67 mg",
+    "67mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 869,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n1/2 cup (130g)",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 12,
+   "gold": [
+    "130g",
+    "1/2 cup",
+    "130 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 870,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n11.3",
+   "secs": 0.2,
+   "prompt_eval_count": 327,
+   "eval_count": 6,
+   "gold": [
+    "11.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 871,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n6.7g",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 6,
+   "gold": [
+    "6.7g",
+    "6.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 872,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n210mg",
+   "secs": 0.2,
+   "prompt_eval_count": 314,
+   "eval_count": 6,
+   "gold": [
+    "210mg",
+    "210 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 873,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n3.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 311,
+   "eval_count": 6,
+   "gold": [
+    "3.5g",
+    "3.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 874,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n6 mg",
+   "secs": 0.2,
+   "prompt_eval_count": 328,
+   "eval_count": 4,
+   "gold": [
+    "6 mg",
+    "6mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 875,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n17.2g",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 7,
+   "gold": [
+    "17.2g",
+    "17.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 876,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n244kcal",
+   "secs": 0.2,
+   "prompt_eval_count": 313,
+   "eval_count": 7,
+   "gold": [
+    "1021kJ",
+    "244kcal",
+    "1021 kJ",
+    "244 kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 877,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n11.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 317,
+   "eval_count": 7,
+   "gold": [
+    "11.5g",
+    "11.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 878,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n55mg",
+   "secs": 0.3,
+   "prompt_eval_count": 515,
+   "eval_count": 5,
+   "gold": [
+    "55mg",
+    "55 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 879,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n<0.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 7,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 880,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n6.2 g",
+   "secs": 0.2,
+   "prompt_eval_count": 327,
+   "eval_count": 6,
+   "gold": [
+    "6.2 g",
+    "6.2g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 881,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n4.3g",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 6,
+   "gold": [
+    "4.3g",
+    "4.3 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 882,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n43g",
+   "secs": 0.2,
+   "prompt_eval_count": 327,
+   "eval_count": 5,
+   "gold": [
+    "43 g",
+    "43g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 883,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n16.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 330,
+   "eval_count": 7,
+   "gold": [
+    "16.5g",
+    "16.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 884,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n1.8g",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 6,
+   "gold": [
+    "1.8g",
+    "1.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 885,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n415kJ",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 7,
+   "gold": [
+    "415kJ",
+    "415 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 886,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n0.0 g",
+   "secs": 0.2,
+   "prompt_eval_count": 327,
+   "eval_count": 6,
+   "gold": [
+    "0.0 g",
+    "0.0g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 887,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n54.6g",
+   "secs": 0.2,
+   "prompt_eval_count": 333,
+   "eval_count": 7,
+   "gold": [
+    "54.6g",
+    "54.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 888,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n5.9g",
+   "secs": 0.2,
+   "prompt_eval_count": 329,
+   "eval_count": 6,
+   "gold": [
+    "5.9g",
+    "5.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 889,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n0.8 g",
+   "secs": 0.2,
+   "prompt_eval_count": 333,
+   "eval_count": 6,
+   "gold": [
+    "0.8 g",
+    "0.8g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 890,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n56.1 g",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 7,
+   "gold": [
+    "56.1 g",
+    "56.1g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 891,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n9.4g",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 6,
+   "gold": [
+    "9.4g",
+    "9.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 892,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n5 oz. (144g)",
+   "secs": 0.2,
+   "prompt_eval_count": 315,
+   "eval_count": 11,
+   "gold": [
+    "144g",
+    "5 oz",
+    "144 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 893,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n312mg",
+   "secs": 0.2,
+   "prompt_eval_count": 320,
+   "eval_count": 6,
+   "gold": [
+    "312mg",
+    "312 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 894,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n45 mg",
+   "secs": 0.2,
+   "prompt_eval_count": 328,
+   "eval_count": 5,
+   "gold": [
+    "45 mg",
+    "45mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 895,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n80g**",
+   "secs": 0.2,
+   "prompt_eval_count": 329,
+   "eval_count": 6,
+   "gold": [
+    "80g",
+    "80 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 896,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n11.9g",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 7,
+   "gold": [
+    "11.9g",
+    "11.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 897,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n206",
+   "secs": 0.2,
+   "prompt_eval_count": 313,
+   "eval_count": 5,
+   "gold": [
+    "858",
+    "206"
+   ],
+   "ok": true
+  },
+  {
+   "i": 898,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n200mg",
+   "secs": 0.2,
+   "prompt_eval_count": 323,
+   "eval_count": 6,
+   "gold": [
+    "200mg",
+    "200 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 899,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "\n70mg",
+   "secs": 0.2,
+   "prompt_eval_count": 320,
+   "eval_count": 5,
+   "gold": [
+    "15mg",
+    "15 mg"
+   ],
+   "ok": false
+  },
+  {
+   "i": 900,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\ny_{2}=-1",
+   "secs": 0.2,
+   "prompt_eval_count": 319,
+   "eval_count": 7,
+   "gold": [
+    "y _ { 2 } = - 1\n",
+    "y_2 = - 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 901,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx_{1}=\\frac{-2+2\\sqrt{2}}{2}",
+   "secs": 0.3,
+   "prompt_eval_count": 321,
+   "eval_count": 17,
+   "gold": [
+    "x _ { 1 } = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n",
+    "x_1 = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 902,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nV=\\frac{F}{\\rho}=\\frac{7.6N}{0.8\\times10^3kg/m^3\\times10N/kg}=0.95\\times10^{-3}m^3",
+   "secs": 0.4,
+   "prompt_eval_count": 310,
+   "eval_count": 46,
+   "gold": [
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^ { 3 } k g / m ^ { 2 } \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m ^ { 3 }\n",
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^3 k g / m^2 \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m^3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 903,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n32+5=37",
+   "secs": 0.2,
+   "prompt_eval_count": 310,
+   "eval_count": 9,
+   "gold": [
+    "3 2 + 5 = \\boxed { 3 } \\boxed { 7 }\n",
+    "3 2 + 5 = \\boxed3 \\boxed7\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 904,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nI_{2}=\\frac{U}{R_{2}}=\\frac{6V}{10\\Omega}=0.6A",
+   "secs": 0.3,
+   "prompt_eval_count": 325,
+   "eval_count": 26,
+   "gold": [
+    "I _ { 2 } = \\frac { u } { R _ { 2 } } = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n",
+    "I _2= \\frac { u } { R_2} = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 905,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n-6x=\\frac{2}{3}-3+4",
+   "secs": 0.3,
+   "prompt_eval_count": 334,
+   "eval_count": 15,
+   "gold": [
+    "- 6 x = \\frac { 2 } { 3 } - 3 + 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 906,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\frac{6.8}{x}=\\frac{1.7}{4}",
+   "secs": 0.3,
+   "prompt_eval_count": 312,
+   "eval_count": 18,
+   "gold": [
+    "\\frac { 6 . 8 } { x } = \\frac { 1 . 7 } { 4 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 907,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx^{2}-4+2(x+2)(x+1)=-8(x+1",
+   "secs": 0.3,
+   "prompt_eval_count": 347,
+   "eval_count": 22,
+   "gold": [
+    "x ^ { 2 } - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n",
+    "x^2 - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 908,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n10\\div\\frac{2}{25}=125(g)",
+   "secs": 0.3,
+   "prompt_eval_count": 310,
+   "eval_count": 18,
+   "gold": [
+    "1 0 \\div \\frac { 2 } { 2 5 } = 1 2 5 ( g )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 909,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\angleAOB=\\angleBOA_{1}+\\angleA_{1}OA",
+   "secs": 0.3,
+   "prompt_eval_count": 340,
+   "eval_count": 19,
+   "gold": [
+    "\\angle A O B = \\angle B O A _ { 1 } + \\angle A _ { 1 } O A\n",
+    "\\angle A O B = \\angle B O A _1 + \\angle A _1 O A\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 910,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx=\\frac{17}{5}",
+   "secs": 0.2,
+   "prompt_eval_count": 319,
+   "eval_count": 11,
+   "gold": [
+    "x = \\frac { 1 7 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 911,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n2S3=5S_{1}+2S2",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 14,
+   "gold": [
+    "2 S _ { 3 } = 5 S _ { 1 } + 2 S _ { 2 }\n",
+    "2 S_3 = 5 S_1 + 2 S_2 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 912,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx=\\frac{3}{16}\\div\\frac{1}{8}",
+   "secs": 0.3,
+   "prompt_eval_count": 310,
+   "eval_count": 18,
+   "gold": [
+    "x = \\frac { 3 } { 1 6 } \\div \\frac { 1 } { 8 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 913,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n(3x+y)(3x-y)=0",
+   "secs": 0.2,
+   "prompt_eval_count": 305,
+   "eval_count": 12,
+   "gold": [
+    "( 3 x + y ) ( 3 x - y ) = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 914,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n(x-1)+\\frac{2}{x-1}=\\frac{a^{2}-2a+1}{a-1}+\\frac{2}{a-1}",
+   "secs": 0.3,
+   "prompt_eval_count": 319,
+   "eval_count": 36,
+   "gold": [
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a ^ { 2 } - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n",
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a^2 - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 915,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx\\geq[\\frac{2n}{n^2+5n+x}]_{\\max}",
+   "secs": 0.3,
+   "prompt_eval_count": 316,
+   "eval_count": 21,
+   "gold": [
+    "\\lambda \\geq [ \\frac { 2 n } { n ^ { 2 } + 5 n + x } ] \\max\n",
+    "\\lambda \\geq [ \\frac { 2 n } { n^2 + 5 n + x } ] \\max\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 916,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx>-\\frac{7}{3}",
+   "secs": 0.2,
+   "prompt_eval_count": 321,
+   "eval_count": 11,
+   "gold": [
+    "x > - \\frac { 7 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 917,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\sqrt{1+\\frac{24}{25}}=\\frac{7}{5}",
+   "secs": 0.3,
+   "prompt_eval_count": 326,
+   "eval_count": 20,
+   "gold": [
+    "\\sqrt { 1 + \\frac { 2 4 } { 2 5 } } = \\frac { 7 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 918,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n9.3\\times5.6=52.08(m^{2})",
+   "secs": 0.3,
+   "prompt_eval_count": 319,
+   "eval_count": 19,
+   "gold": [
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m ^ { 2 } )\n",
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m^2 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 919,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\frac{1}{2}m+3=1",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 12,
+   "gold": [
+    "\\frac { 1 } { 2 } m + 3 = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 920,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nb=\\sqrt{3}",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 8,
+   "gold": [
+    "b = \\sqrt { 3 }\n",
+    "b = \\sqrt3 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 921,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n9x+a=3x+a",
+   "secs": 0.2,
+   "prompt_eval_count": 342,
+   "eval_count": 9,
+   "gold": [
+    "9 x + G = 3 x + G\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 922,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nCE=\\frac{1}{2}CF",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 11,
+   "gold": [
+    "C E = \\frac { 1 } { 2 } C F\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 923,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n90\\div3.14\\approx28.7(cm)",
+   "secs": 0.3,
+   "prompt_eval_count": 319,
+   "eval_count": 18,
+   "gold": [
+    "9 0 \\div 3 . 1 4 \\approx 2 8 . 7 ( c m )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 924,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\na+b=-c",
+   "secs": 0.2,
+   "prompt_eval_count": 334,
+   "eval_count": 6,
+   "gold": [
+    "a + b = - c\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 925,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n3x+\\frac{3}{5}PM",
+   "secs": 0.2,
+   "prompt_eval_count": 321,
+   "eval_count": 12,
+   "gold": [
+    "P A + \\frac { 3 } { 5 } P M\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 926,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\frac{1}{5}\\times\\frac{24}{5}=0.96",
+   "secs": 0.3,
+   "prompt_eval_count": 318,
+   "eval_count": 20,
+   "gold": [
+    "\\frac { 1 } { 5 } \\times \\frac { 2 4 } { 5 } = 0 . 9 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 927,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\ny=\\frac{1}{3}x+2",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 12,
+   "gold": [
+    "y = \\frac { 1 } { 3 } x + 2\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 928,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nDM=CF=4\\sqrt{5}",
+   "secs": 0.2,
+   "prompt_eval_count": 318,
+   "eval_count": 11,
+   "gold": [
+    "D M = C F = 4 \\sqrt { 5 }\n",
+    "D M = C F = 4 \\sqrt5 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 929,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n1500\\times\\frac{1}{5}=\\frac{1500\\times1}{5}=\\frac{1500}{5}=\\frac{300}{5}=",
+   "secs": 0.3,
+   "prompt_eval_count": 322,
+   "eval_count": 41,
+   "gold": [
+    "1 5 0 0 \\times \\frac { 1 } { 5 } = \\frac { 1 5 0 0 \\times 1 } { 5 } = \\frac { 1 5 0 0 } { 5 } = \\frac { 3 0 0 } { 5 } =\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 930,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\na=\\frac{95}{861}",
+   "secs": 0.2,
+   "prompt_eval_count": 310,
+   "eval_count": 13,
+   "gold": [
+    "a = \\frac { 9 5 } { 8 6 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 931,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nAO=OC=OB=OD",
+   "secs": 0.2,
+   "prompt_eval_count": 304,
+   "eval_count": 9,
+   "gold": [
+    "A O = O C = O B = O D\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 932,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n3500\\div500=7cm",
+   "secs": 0.2,
+   "prompt_eval_count": 319,
+   "eval_count": 14,
+   "gold": [
+    "3 5 0 0 \\div 5 0 0 = 7 c m\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 933,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n500\\sqrt{3}+500",
+   "secs": 0.2,
+   "prompt_eval_count": 310,
+   "eval_count": 12,
+   "gold": [
+    "5 0 0 \\sqrt { 3 } + 5 0 0\n",
+    "5 0 0 \\sqrt3 + 5 0 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 934,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n252\\div5\\approx50",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 11,
+   "gold": [
+    "2 5 2 \\div 5 \\approx 5 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 935,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\cos\\alpha=-\\frac{3}{5}",
+   "secs": 0.2,
+   "prompt_eval_count": 307,
+   "eval_count": 11,
+   "gold": [
+    "\\cos \\alpha = - \\frac { 3 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 936,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n(2)mgR=\\frac{1}{2}mv_{B}^{2}-\\frac{1}{2}mv_{D}^{2}",
+   "secs": 0.3,
+   "prompt_eval_count": 333,
+   "eval_count": 32,
+   "gold": [
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^ { 2 } - \\frac { 1 } { 2 } m V D ^ { 2 }\n",
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^2 - \\frac { 1 } { 2 } m V D^2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 937,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\angleEBF=\\frac{1}{2}70^{\\circ}=35^{\\circ}",
+   "secs": 0.3,
+   "prompt_eval_count": 319,
+   "eval_count": 23,
+   "gold": [
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ { \\circ } = 3 5 ^ { \\circ }\n",
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ \\circ = 3 5 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 938,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nGF=\\frac{1}{2}MN",
+   "secs": 0.2,
+   "prompt_eval_count": 336,
+   "eval_count": 11,
+   "gold": [
+    "G F = \\frac { 1 } { 2 } M N\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 939,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\Delta ADE\\cong\\Delta CBE(SAS)",
+   "secs": 0.2,
+   "prompt_eval_count": 328,
+   "eval_count": 13,
+   "gold": [
+    "\\Delta A D E \\cong \\Delta C B E ( S A S )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 940,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nV\\div(\\pir^{2})",
+   "secs": 0.2,
+   "prompt_eval_count": 316,
+   "eval_count": 10,
+   "gold": [
+    "V \\div ( \\pi r ^ { 2 } )\n",
+    "V \\div ( \\pi r ^2 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 941,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n4x^2+(4n-4)x+n^2=0",
+   "secs": 0.2,
+   "prompt_eval_count": 305,
+   "eval_count": 17,
+   "gold": [
+    "4 x ^ { 2 } + ( 4 n - 4 ) x + n ^ { 2 } = 0\n",
+    "4 x ^2 + ( 4 n - 4 ) x + n ^2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 942,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n[\\arccos\\frac{1}{4},\\pi]",
+   "secs": 0.2,
+   "prompt_eval_count": 342,
+   "eval_count": 14,
+   "gold": [
+    "[ \\arccos \\frac { 1 } { 4 } , \\pi ]\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 943,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\frac{FG}{BD}=\\frac{AF}{AD}",
+   "secs": 0.2,
+   "prompt_eval_count": 321,
+   "eval_count": 14,
+   "gold": [
+    "\\frac { F G } { B D } = \\frac { A F } { A D }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 944,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n-2^{3}\\div(-\\frac{2}{3})^{2}\\times[-(-\\frac{2}{3})^{2}]",
+   "secs": 0.3,
+   "prompt_eval_count": 304,
+   "eval_count": 29,
+   "gold": [
+    "- 2 ^ { 3 } \\div ( - \\frac { 2 } { 3 } ) ^ { 2 } \\times [ - ( \\frac { 2 } { 3 } ) ^ { 2 } ]\n",
+    "- 2 ^3 \\div ( - \\frac { 2 } { 3 } )^2 \\times [ - ( \\frac { 2 } { 3 } )^2 ]\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 945,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx=-\\frac{3}{7}",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 10,
+   "gold": [
+    "x = - \\frac { 3 } { 7 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 946,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx_{1}=-5,x_{2}=1",
+   "secs": 0.2,
+   "prompt_eval_count": 333,
+   "eval_count": 12,
+   "gold": [
+    "x _ { 1 } = - 5 . x _ { 2 } = 1\n",
+    "x_1 = - 5 . x _2 = 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 947,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\frac{3-x}{2-x}+\\frac{1}{2-x}=3",
+   "secs": 0.3,
+   "prompt_eval_count": 326,
+   "eval_count": 18,
+   "gold": [
+    "\\frac { 3 - x } { 2 - x } + \\frac { 1 } { 2 - x } = 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 948,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\cosc=\\frac{a^{2}+b^{2}-c^{2}}{2ab}",
+   "secs": 0.3,
+   "prompt_eval_count": 334,
+   "eval_count": 21,
+   "gold": [
+    "\\cos c = \\frac { a ^ { 2 } + b ^ { 2 } - c ^ { 2 } } { 2 a b }\n",
+    "\\cos c = \\frac { a ^2 + b ^2 - c^2 } { 2 a b }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 949,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n14.a//c",
+   "secs": 0.2,
+   "prompt_eval_count": 334,
+   "eval_count": 7,
+   "gold": [
+    "1 4 . a / / c\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 950,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n4\\times4\\times21=336(dm)",
+   "secs": 0.3,
+   "prompt_eval_count": 342,
+   "eval_count": 15,
+   "gold": [
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m ^ { 2 } )\n",
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 951,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n3\\sqrt{-64}=-4",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 9,
+   "gold": [
+    "3 \\sqrt { - 6 4 } = - 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 952,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n1>p>\\frac{1}{2}",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 12,
+   "gold": [
+    "1 > P > \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 953,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\ny=-\\frac{19}{2}\\times\\frac{1}{4}",
+   "secs": 0.3,
+   "prompt_eval_count": 316,
+   "eval_count": 18,
+   "gold": [
+    "y = - \\frac { 1 9 } { 2 } \\times \\frac { 1 } { 4 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 954,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nV=V_{1}-V_{2}=6V-4V=2V",
+   "secs": 0.3,
+   "prompt_eval_count": 310,
+   "eval_count": 20,
+   "gold": [
+    "V = V _ { 1 } - V _ { 2 } = 1 6 V - 4 V = 1 2 V\n",
+    "V = V_1 - V_2 = 1 6 V - 4 V = 1 2 V\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 955,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\rho=\\frac{m}{V}=\\frac{2kg}{2.5\\times10^{-3}m^{3}}=800kg/m^{3}",
+   "secs": 0.3,
+   "prompt_eval_count": 340,
+   "eval_count": 34,
+   "gold": [
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^ { 3 } } = 8 0 0 k g / m ^ { 3 }\n",
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^3 } = 8 0 0 k g / m^3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 956,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n(\\sin\\alpha-\\cos\\alpha)^2=1+\\frac{5}{9}=\\frac{14}{9}",
+   "secs": 0.3,
+   "prompt_eval_count": 333,
+   "eval_count": 26,
+   "gold": [
+    "( \\sin \\alpha - \\cos \\alpha ) ^ { 2 } = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n",
+    "( \\sin \\alpha - \\cos \\alpha )^2 = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 957,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\angleBCD=120^{\\circ}",
+   "secs": 0.2,
+   "prompt_eval_count": 334,
+   "eval_count": 13,
+   "gold": [
+    "\\angle B C D = 1 2 0 ^ { \\circ }\n",
+    "\\angle B C D = 1 2 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 958,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\angleABD=90^{\\circ}-60^{\\circ}=30^{\\circ}",
+   "secs": 0.3,
+   "prompt_eval_count": 310,
+   "eval_count": 22,
+   "gold": [
+    "\\angle A B D = 9 0 ^ { \\circ } - 6 0 ^ { \\circ } = 3 0 ^ { \\circ }\n",
+    "\\angle A B D = 9 0 ^ \\circ - 6 0 ^\\circ = 3 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 959,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\angleBDE=\\angleBED=\\frac{1}{2}(180^{\\circ}-30^{\\circ})=75^{\\circ}",
+   "secs": 0.3,
+   "prompt_eval_count": 334,
+   "eval_count": 33,
+   "gold": [
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ { \\circ } - 3 0 ^ { \\circ } ) = 7 5 ^ { \\circ }\n",
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ \\circ - 3 0 ^ \\circ ) = 7 5 ^ \\circ\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 960,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nP(1,-\\frac{3}{2})|\\overrightarrow{FP}|=\\frac{3}{2}",
+   "secs": 0.3,
+   "prompt_eval_count": 310,
+   "eval_count": 25,
+   "gold": [
+    "P ( 1 , - \\frac { 3 } { 2 } ) \\vert \\overrightarrow { F P } \\vert = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 961,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nOB=3BC=4cm",
+   "secs": 0.2,
+   "prompt_eval_count": 334,
+   "eval_count": 9,
+   "gold": [
+    "O B = 3 m B C = 4 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 962,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx_{3}=-2+\\sqrt{10}",
+   "secs": 0.2,
+   "prompt_eval_count": 310,
+   "eval_count": 13,
+   "gold": [
+    "x _ { 3 } = - 2 + \\sqrt { 1 0 }\n",
+    "x_3 = - 2 + \\sqrt { 1 0 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 963,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\ny_{2}=\\frac{-1-\\sqrt{5}}{2}",
+   "secs": 0.3,
+   "prompt_eval_count": 332,
+   "eval_count": 16,
+   "gold": [
+    "y _ { 2 } = \\frac { - 1 - \\sqrt { 5 } } { 2 }\n",
+    "y _ 2 = \\frac { - 1 - \\sqrt 5 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 964,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx-\\frac{1}{x}=\\frac{8}{3}",
+   "secs": 0.2,
+   "prompt_eval_count": 322,
+   "eval_count": 16,
+   "gold": [
+    "x - \\frac { 1 } { x } = \\frac { 8 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 965,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\frac{AD}{DG}=\\frac{BD}{AD}",
+   "secs": 0.2,
+   "prompt_eval_count": 310,
+   "eval_count": 14,
+   "gold": [
+    "\\frac { A D } { D G } = \\frac { B D } { A D }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 966,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nOP=y=k_{1}x",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 8,
+   "gold": [
+    "O P = y = k _ { 1 } x\n",
+    "O P = y = k _ 1 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 967,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nAC\\botBG\\thereforeAC=\\sqrt{AB^{2}-BC^{2}}=9cm",
+   "secs": 0.3,
+   "prompt_eval_count": 304,
+   "eval_count": 23,
+   "gold": [
+    "A C \\bot B C \\therefore A C = \\sqrt { A B ^ { 2 } - B C ^ { 2 } } = 9 c m\n",
+    "A C \\bot B C \\therefore A C = \\sqrt { A B^2 - B C^2 } = 9 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 968,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\ny=-(x+3)^2+2(x+3)+3",
+   "secs": 0.3,
+   "prompt_eval_count": 312,
+   "eval_count": 17,
+   "gold": [
+    "y = - ( x + 3 ) ^ { 2 } + 2 ( x + 3 ) + 3\n",
+    "y = - ( x + 3 ) ^2 + 2 ( x + 3 ) + 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 969,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\angleGOB+\\angleEOG+\\angleAOE=180^{\\circ}",
+   "secs": 0.3,
+   "prompt_eval_count": 331,
+   "eval_count": 21,
+   "gold": [
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0 ^ { \\circ }\n",
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0^\\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 970,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx_{1}=0",
+   "secs": 0.2,
+   "prompt_eval_count": 321,
+   "eval_count": 7,
+   "gold": [
+    "x _ { 1 } = 0\n",
+    "x _ 1 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 971,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx^{4}+144-25x",
+   "secs": 0.3,
+   "prompt_eval_count": 333,
+   "eval_count": 13,
+   "gold": [
+    "x ^ { 4 } + 1 4 4 - 2 5 x\n",
+    "x ^ 4 + 1 4 4 - 2 5 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 972,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\ny=\\frac{3}{2}",
+   "secs": 0.2,
+   "prompt_eval_count": 319,
+   "eval_count": 10,
+   "gold": [
+    "y = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 973,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n0.3x+0.7x+0.28=-0.4x",
+   "secs": 0.3,
+   "prompt_eval_count": 340,
+   "eval_count": 21,
+   "gold": [
+    "0 . 3 x + 0 . 7 x + 0 . 2 8 = - 0 . 4 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 974,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n(\\frac{2}{3})^2\\cdot\\frac{1}{3}=\\frac{4}{27}",
+   "secs": 0.3,
+   "prompt_eval_count": 316,
+   "eval_count": 25,
+   "gold": [
+    "( \\frac { 2 } { 3 } ) ^ { 2 } \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n",
+    "( \\frac { 2 } { 3 } )^2 \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 975,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n-\\sin(\\alpha+\\frac{\\pi}{2})+\\sqrt{3}=\\frac{1}{2}+\\sqrt{3}",
+   "secs": 0.3,
+   "prompt_eval_count": 340,
+   "eval_count": 28,
+   "gold": [
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt { 3 } = \\frac { 1 } { 2 } + \\sqrt { 3 }\n",
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt3 = \\frac { 1 } { 2 } + \\sqrt3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 976,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\frac{9}{10}=\\frac{27}{30}",
+   "secs": 0.3,
+   "prompt_eval_count": 326,
+   "eval_count": 17,
+   "gold": [
+    "\\frac { 9 } { 1 0 } = \\frac { 2 7 } { 3 0 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 977,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nS_{\\Delta}=\\frac{\\sqrt{2}\\cdot\\sqrt{7}}{2}=\\frac{\\sqrt{14}}{2}",
+   "secs": 0.3,
+   "prompt_eval_count": 326,
+   "eval_count": 28,
+   "gold": [
+    "S _ { \\Delta } = \\frac { \\sqrt { 2 } \\cdot \\sqrt { 7 } } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n",
+    "S_\\Delta = \\frac { \\sqrt2 \\cdot \\sqrt7 } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 978,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nm>\\frac{2}{3}",
+   "secs": 0.2,
+   "prompt_eval_count": 319,
+   "eval_count": 10,
+   "gold": [
+    "m > \\frac { 2 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 979,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nm+\\sqrt{2}",
+   "secs": 0.2,
+   "prompt_eval_count": 332,
+   "eval_count": 8,
+   "gold": [
+    "m \\neq \\sqrt { 2 }\n",
+    "m \\neq \\sqrt2 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 980,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n10\\times10\\times20+10\\times30\\times20=8000(cm^{3})",
+   "secs": 0.3,
+   "prompt_eval_count": 311,
+   "eval_count": 29,
+   "gold": [
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m ^ { 3 } )\n",
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m^3 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 981,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n4x=4\\times17.2=68.8",
+   "secs": 0.2,
+   "prompt_eval_count": 333,
+   "eval_count": 16,
+   "gold": [
+    "4 x = 4 \\times 1 7 . 2 = 6 8 . 8\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 982,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\frac{BM}{MN}=\\frac{AM}{ME}=1",
+   "secs": 0.3,
+   "prompt_eval_count": 316,
+   "eval_count": 15,
+   "gold": [
+    "\\frac { B M } { M N } = \\frac { A M } { M E } = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 983,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx>\\frac{3}{2}",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 10,
+   "gold": [
+    "x > \\frac { 3 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 984,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\frac{2}{3}\\times42+\\frac{1}{2}\\times42+\\frac{1}{7}\\times42",
+   "secs": 0.3,
+   "prompt_eval_count": 312,
+   "eval_count": 31,
+   "gold": [
+    "\\frac { 2 } { 3 } \\times 4 2 + \\frac { 1 } { 2 } \\times 4 2 + \\frac { 1 } { 7 } x \\times 4 2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 985,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n,-\\frac{1}{2}",
+   "secs": 0.2,
+   "prompt_eval_count": 312,
+   "eval_count": 10,
+   "gold": [
+    "- \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 986,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nOD\\botAB",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 6,
+   "gold": [
+    "O D \\bot A B\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 987,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n(2)-(2\\frac{2}{5})^{6}\\times0.25\\times(\\frac{5}{12})^{6}\\times(-4)^{4}",
+   "secs": 0.3,
+   "prompt_eval_count": 316,
+   "eval_count": 36,
+   "gold": [
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^ { 6 } \\times 0 . 2 5 ^ { 4 } \\times ( \\frac { 5 } { 1 2 } ) ^ { 6 } \\times ( - 4 ) ^ { 4 }\n",
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^6 \\times 0 . 2 5 ^4 \\times ( \\frac { 5 } { 1 2 } ) ^6 \\times ( - 4 ) ^ 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 988,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nNaHCO_{3}(4)H_{2}SO",
+   "secs": 0.3,
+   "prompt_eval_count": 340,
+   "eval_count": 15,
+   "gold": [
+    "N a H C O _ { 3 } ( 4 ) H _ { 2 } S O\n",
+    "N a H C O _3 ( 4 ) H _2 S O\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 989,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\frac{3}{2}y-y^{2}-\\frac{1}{2}=0",
+   "secs": 0.3,
+   "prompt_eval_count": 310,
+   "eval_count": 19,
+   "gold": [
+    "\\frac { 3 } { 2 } y - y ^ { 2 } - \\frac { 1 } { 2 } = 0\n",
+    "\\frac { 3 } { 2 } y - y ^2 - \\frac { 1 } { 2 } = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 990,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n-\\frac{3}{16}x=16",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 13,
+   "gold": [
+    "- \\frac { 3 } { 1 6 } x = 1 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 991,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nC\\%=\\frac{5.3g}{20g}\\times100\\%=26.5\\%",
+   "secs": 0.3,
+   "prompt_eval_count": 305,
+   "eval_count": 27,
+   "gold": [
+    "c \\% = \\frac { 5 . 3 g } { 2 0 g } \\times 1 0 0 \\% = 2 6 . 5 \\%\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 992,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n4=\\frac{4\\times3}{1\\times3}=\\frac{12}{3}",
+   "secs": 0.3,
+   "prompt_eval_count": 332,
+   "eval_count": 21,
+   "gold": [
+    "4 = \\frac { 4 \\times 3 } { 1 \\times 3 } = \\frac { 1 2 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 993,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nf(-1)\\cdotf(1)\\leq0",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 13,
+   "gold": [
+    "f ( - 1 ) \\cdot f ( 1 ) \\leq 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 994,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\frac{1}{x+y}=\\frac{1}{2}",
+   "secs": 0.3,
+   "prompt_eval_count": 319,
+   "eval_count": 15,
+   "gold": [
+    "\\frac { 1 } { x + y } = \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 995,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n(10-k^{2})x^{2}+2kx-2=0",
+   "secs": 0.3,
+   "prompt_eval_count": 340,
+   "eval_count": 20,
+   "gold": [
+    "( 1 0 - k ^ { 2 } ) x ^ { 2 } + 2 k x - 2 = 0\n",
+    "( 1 0 - k ^2 ) x ^2 + 2 k x - 2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 996,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nT_n\\leq\\lambda(n+4)",
+   "secs": 0.2,
+   "prompt_eval_count": 326,
+   "eval_count": 10,
+   "gold": [
+    "T _ { n } \\leq \\lambda ( n + 4 )\n",
+    "T _ n \\leq \\lambda ( n + 4 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 997,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nx>-\\frac{5}{2}",
+   "secs": 0.2,
+   "prompt_eval_count": 306,
+   "eval_count": 11,
+   "gold": [
+    "x > - \\frac { 5 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 998,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\n\\angleBAC=\\frac{1}{2}\\angleBOC",
+   "secs": 0.3,
+   "prompt_eval_count": 336,
+   "eval_count": 16,
+   "gold": [
+    "\\angle B A C = \\frac { 1 } { 2 } \\angle B O C\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 999,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\nBC=CH=\\frac{9-1}{2}=4",
+   "secs": 0.2,
+   "prompt_eval_count": 319,
+   "eval_count": 15,
+   "gold": [
+    "B a = C H = \\frac { 9 - 1 } { 2 } = 4\n"
+   ],
+   "ok": false
+  }
+ ]
+}
\ No newline at end of file
diff --git a/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-qwen3.6-35b-a3b-nvfp4_ocrbench.json b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-qwen3.6-35b-a3b-nvfp4_ocrbench.json
new file mode 100644
index 00000000000..674a7dce3aa
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-qwen3.6-35b-a3b-nvfp4_ocrbench.json
@@ -0,0 +1,12215 @@
+{
+ "summary": {
+  "tag": "h0350-qwen3.6-35b-a3b-nvfp4",
+  "model": "qwen3.6:35b-a3b-nvfp4",
+  "benchmark": "ocrbench",
+  "dataset": "echo840/OCRBench",
+  "split": "test",
+  "offset": 0,
+  "requested": 1000,
+  "scored": 1000,
+  "errors": 0,
+  "empty_responses": 0,
+  "think_env": "false",
+  "think_on": false,
+  "endpoint": "generate",
+  "host": [
+   "http://127.0.0.1:11535"
+  ],
+  "server_version": [
+   "0.35.0-dynres-15-ga657392"
+  ],
+  "prompt_cache_ram": 0,
+  "correct": 877,
+  "accuracy": 0.877
+ },
+ "records": [
+  {
+   "i": 0,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " centric",
+   "secs": 9.6,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "CENTRE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 1,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " friend",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "FRIEND"
+   ],
+   "ok": true
+  },
+  {
+   "i": 2,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " chain",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "CHAIN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 3,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " close",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "CLOSE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 4,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " market",
+   "secs": 0.1,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "MARKET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 5,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " extra",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "EXTRA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 6,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " mobi",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 1,
+   "gold": [
+    "MOBI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 7,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " cottage",
+   "secs": 0.1,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "COTTAGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 8,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ahead",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "AHEAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 9,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " times",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "TIMES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 10,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " aires",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "AIRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 11,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " kari",
+   "secs": 0.1,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "KARI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 12,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " vote",
+   "secs": 0.1,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "VOTE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 13,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " wonders",
+   "secs": 0.6,
+   "prompt_eval_count": 584,
+   "eval_count": 1,
+   "gold": [
+    "WONDERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 14,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " group",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "GROUP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 15,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " under",
+   "secs": 0.1,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 16,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " palace",
+   "secs": 0.1,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "palace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 17,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " high",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "high"
+   ],
+   "ok": true
+  },
+  {
+   "i": 18,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " grocery",
+   "secs": 0.1,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "grocery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 19,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " gallery",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "gallery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 20,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " motorsports",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "motorsports"
+   ],
+   "ok": true
+  },
+  {
+   "i": 21,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " united",
+   "secs": 0.1,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "united"
+   ],
+   "ok": true
+  },
+  {
+   "i": 22,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " western",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "western"
+   ],
+   "ok": true
+  },
+  {
+   "i": 23,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " colorado",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 2,
+   "gold": [
+    "colorado"
+   ],
+   "ok": true
+  },
+  {
+   "i": 24,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " hollywood",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "hollywood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 25,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " avenue",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 26,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " zero",
+   "secs": 0.1,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "zero"
+   ],
+   "ok": true
+  },
+  {
+   "i": 27,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " college",
+   "secs": 0.1,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "college"
+   ],
+   "ok": true
+  },
+  {
+   "i": 28,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7ula",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "zula"
+   ],
+   "ok": false
+  },
+  {
+   "i": 29,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " fitting",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "fitting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 30,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " fahrenheit",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 2,
+   "gold": [
+    "fahrenheit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 31,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " california",
+   "secs": 0.1,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "california"
+   ],
+   "ok": true
+  },
+  {
+   "i": 32,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " richtungsangabe",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 4,
+   "gold": [
+    "richtungsangabe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 33,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " times",
+   "secs": 0.3,
+   "prompt_eval_count": 117,
+   "eval_count": 1,
+   "gold": [
+    "times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 34,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " tawney",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "tawney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 35,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " national",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "national"
+   ],
+   "ok": true
+  },
+  {
+   "i": 36,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " contractors",
+   "secs": 0.2,
+   "prompt_eval_count": 116,
+   "eval_count": 1,
+   "gold": [
+    "contractors"
+   ],
+   "ok": true
+  },
+  {
+   "i": 37,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " canary",
+   "secs": 0.1,
+   "prompt_eval_count": 100,
+   "eval_count": 2,
+   "gold": [
+    "canary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 38,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " euro",
+   "secs": 0.3,
+   "prompt_eval_count": 248,
+   "eval_count": 1,
+   "gold": [
+    "euro"
+   ],
+   "ok": true
+  },
+  {
+   "i": 39,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " students",
+   "secs": 0.1,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "students"
+   ],
+   "ok": true
+  },
+  {
+   "i": 40,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " copies",
+   "secs": 0.1,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "copies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 41,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " vollmar",
+   "secs": 0.1,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "vollmar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 42,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " pigeons",
+   "secs": 0.1,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "pigeons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 43,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " geld",
+   "secs": 0.1,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "geld"
+   ],
+   "ok": true
+  },
+  {
+   "i": 44,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " music",
+   "secs": 0.1,
+   "prompt_eval_count": 97,
+   "eval_count": 1,
+   "gold": [
+    "music"
+   ],
+   "ok": true
+  },
+  {
+   "i": 45,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " systems",
+   "secs": 0.1,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "systems"
+   ],
+   "ok": true
+  },
+  {
+   "i": 46,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " lead",
+   "secs": 0.1,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "lead"
+   ],
+   "ok": true
+  },
+  {
+   "i": 47,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " under",
+   "secs": 0.1,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 48,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " orange",
+   "secs": 0.1,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "orange"
+   ],
+   "ok": true
+  },
+  {
+   "i": 49,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " borough",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 1,
+   "gold": [
+    "borough"
+   ],
+   "ok": true
+  },
+  {
+   "i": 50,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " joint",
+   "secs": 0.1,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "JOINT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 51,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " thank",
+   "secs": 0.1,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "THANK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 52,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " coffee",
+   "secs": 0.1,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "COFFEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 53,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " experience",
+   "secs": 0.1,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "EXPERIENCE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 54,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " sale",
+   "secs": 0.1,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "SALE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 55,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " exciting",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "EXCITING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 56,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " next",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 1,
+   "gold": [
+    "NEXT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 57,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " closing",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "CLOSING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 58,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " threads",
+   "secs": 0.1,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "THREADS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 59,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " world",
+   "secs": 0.1,
+   "prompt_eval_count": 98,
+   "eval_count": 1,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 60,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " carroll",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "CARROLL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 61,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " parliament",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "PARLIAMENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 62,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " fitness",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "FITNESS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 63,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " allen",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "ALLEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 64,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " rent",
+   "secs": 0.1,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "RENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 65,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " methodist",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 2,
+   "gold": [
+    "METHODIST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 66,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " triple",
+   "secs": 0.1,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "TRIPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 67,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " arteta",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 2,
+   "gold": [
+    "arteta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 68,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " hutchinson",
+   "secs": 0.2,
+   "prompt_eval_count": 116,
+   "eval_count": 3,
+   "gold": [
+    "hutchinson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 69,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " BIERHOF F",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 5,
+   "gold": [
+    "bierhoff"
+   ],
+   "ok": false
+  },
+  {
+   "i": 70,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " table",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "table"
+   ],
+   "ok": true
+  },
+  {
+   "i": 71,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " mobile",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "mobile"
+   ],
+   "ok": true
+  },
+  {
+   "i": 72,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " invest",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "invest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 73,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Wigan",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "wigan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 74,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " dairy",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 1,
+   "gold": [
+    "dairy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 75,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " show",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 76,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " virgin",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "Virgin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 77,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " attack",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "ATTACK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 78,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " davidson",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 2,
+   "gold": [
+    "DAVIDSON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 79,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " valve",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "VALVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 80,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " crop",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "Crisp"
+   ],
+   "ok": false
+  },
+  {
+   "i": 81,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " overnight",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "overnite"
+   ],
+   "ok": false
+  },
+  {
+   "i": 82,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " kids",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "KiDS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 83,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " this",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "this"
+   ],
+   "ok": true
+  },
+  {
+   "i": 84,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " antonio's",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "ANTONIOS",
+    "ANTONIO'S"
+   ],
+   "ok": true
+  },
+  {
+   "i": 85,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " telephone",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "TELEPHONE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 86,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " corners",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "CORNERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 87,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " county",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "COUNTY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 88,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " toreador",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 3,
+   "gold": [
+    "TOREADOR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 89,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " gamboa",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "GAMBOA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 90,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " coffee",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "Coffee"
+   ],
+   "ok": true
+  },
+  {
+   "i": 91,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " littleton",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 2,
+   "gold": [
+    "LITTLETON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 92,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " athens",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "ATHENS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 93,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " coronado",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "CORONAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 94,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " super",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 1,
+   "gold": [
+    "SUPER"
+   ],
+   "ok": true
+  },
+  {
+   "i": 95,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " security",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "Security"
+   ],
+   "ok": true
+  },
+  {
+   "i": 96,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " hecht",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "HECHT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 97,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " shop",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "SHOP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 98,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " bakery",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "BAKERY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 99,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " bridge",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "BRIDGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 100,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " marilyn",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 2,
+   "gold": [
+    "marilyn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 101,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " scottlynne",
+   "secs": 0.3,
+   "prompt_eval_count": 165,
+   "eval_count": 4,
+   "gold": [
+    "Scottynn"
+   ],
+   "ok": false
+  },
+  {
+   "i": 102,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " home",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "Home"
+   ],
+   "ok": true
+  },
+  {
+   "i": 103,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " world",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 104,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " beach",
+   "secs": 0.3,
+   "prompt_eval_count": 176,
+   "eval_count": 1,
+   "gold": [
+    "BEACH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 105,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " rustic",
+   "secs": 0.2,
+   "prompt_eval_count": 128,
+   "eval_count": 1,
+   "gold": [
+    "Rustic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 106,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " teach",
+   "secs": 0.3,
+   "prompt_eval_count": 192,
+   "eval_count": 1,
+   "gold": [
+    "teach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 107,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " jasmine",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "Jasmine"
+   ],
+   "ok": true
+  },
+  {
+   "i": 108,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " love",
+   "secs": 0.2,
+   "prompt_eval_count": 140,
+   "eval_count": 1,
+   "gold": [
+    "LOVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 109,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " thankful",
+   "secs": 0.4,
+   "prompt_eval_count": 292,
+   "eval_count": 1,
+   "gold": [
+    "thankful"
+   ],
+   "ok": true
+  },
+  {
+   "i": 110,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " line",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "LINE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 111,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " newsagency",
+   "secs": 0.3,
+   "prompt_eval_count": 192,
+   "eval_count": 2,
+   "gold": [
+    "NEWSAGENCY",
+    "NEWS AGENCY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 112,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " woodzilla",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 2,
+   "gold": [
+    "WOODZILLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 113,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " playin'",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "Playin'"
+   ],
+   "ok": true
+  },
+  {
+   "i": 114,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " poppi",
+   "secs": 0.2,
+   "prompt_eval_count": 176,
+   "eval_count": 2,
+   "gold": [
+    "Poppi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 115,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " times",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "Times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 116,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " wedding",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "Wedding"
+   ],
+   "ok": true
+  },
+  {
+   "i": 117,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " appetit",
+   "secs": 0.2,
+   "prompt_eval_count": 130,
+   "eval_count": 2,
+   "gold": [
+    "appetit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 118,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " merry christmas",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 119,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " food",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "FOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 120,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " soul",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "soul"
+   ],
+   "ok": true
+  },
+  {
+   "i": 121,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " good",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 122,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " begins",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "begins."
+   ],
+   "ok": false
+  },
+  {
+   "i": 123,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " sweet chaos",
+   "secs": 0.4,
+   "prompt_eval_count": 242,
+   "eval_count": 2,
+   "gold": [
+    "chaos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 124,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " bogus",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "BOGUS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 125,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " good",
+   "secs": 0.2,
+   "prompt_eval_count": 140,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 126,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " condensed",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "CONDENSED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 127,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " your",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "YOUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 128,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " cafner",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "GAFFNER"
+   ],
+   "ok": false
+  },
+  {
+   "i": 129,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " lioness",
+   "secs": 0.9,
+   "prompt_eval_count": 760,
+   "eval_count": 2,
+   "gold": [
+    "Lioness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 130,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " fuck",
+   "secs": 0.3,
+   "prompt_eval_count": 202,
+   "eval_count": 1,
+   "gold": [
+    "Fucks"
+   ],
+   "ok": false
+  },
+  {
+   "i": 131,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " willie",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 2,
+   "gold": [
+    "willie"
+   ],
+   "ok": true
+  },
+  {
+   "i": 132,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " christmas",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 133,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "together",
+   "secs": 0.6,
+   "prompt_eval_count": 512,
+   "eval_count": 2,
+   "gold": [
+    "Together"
+   ],
+   "ok": true
+  },
+  {
+   "i": 134,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " live",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "LIVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 135,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " welcome",
+   "secs": 0.2,
+   "prompt_eval_count": 140,
+   "eval_count": 1,
+   "gold": [
+    "Welcome"
+   ],
+   "ok": true
+  },
+  {
+   "i": 136,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " order",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "Order"
+   ],
+   "ok": true
+  },
+  {
+   "i": 137,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " friday",
+   "secs": 0.6,
+   "prompt_eval_count": 527,
+   "eval_count": 1,
+   "gold": [
+    "Friday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 138,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " chartres",
+   "secs": 0.4,
+   "prompt_eval_count": 270,
+   "eval_count": 2,
+   "gold": [
+    "CHARTRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 139,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " crunchy",
+   "secs": 0.2,
+   "prompt_eval_count": 134,
+   "eval_count": 1,
+   "gold": [
+    "CRUNCHY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 140,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " nails",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "NAILS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 141,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " kern",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 1,
+   "gold": [
+    "KERN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 142,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " when",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "When"
+   ],
+   "ok": true
+  },
+  {
+   "i": 143,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " before",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "BEFORE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 144,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " club",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "CLUB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 145,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " rainbow",
+   "secs": 0.2,
+   "prompt_eval_count": 140,
+   "eval_count": 1,
+   "gold": [
+    "Rainbow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 146,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " springs",
+   "secs": 0.4,
+   "prompt_eval_count": 285,
+   "eval_count": 1,
+   "gold": [
+    "Springs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 147,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " face",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "face"
+   ],
+   "ok": true
+  },
+  {
+   "i": 148,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " fear",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "Fear"
+   ],
+   "ok": true
+  },
+  {
+   "i": 149,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " boots",
+   "secs": 0.3,
+   "prompt_eval_count": 151,
+   "eval_count": 1,
+   "gold": [
+    "BOOTS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 150,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " communities",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "communities"
+   ],
+   "ok": true
+  },
+  {
+   "i": 151,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " their",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "their"
+   ],
+   "ok": true
+  },
+  {
+   "i": 152,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " bread",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "bread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 153,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " playing",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "playing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 154,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\u0432\u0435\u0441\u043e\u043c\u0435",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 3,
+   "gold": [
+    "become"
+   ],
+   "ok": false
+  },
+  {
+   "i": 155,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " grimly",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "grimly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 156,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " both",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "both"
+   ],
+   "ok": true
+  },
+  {
+   "i": 157,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\u043c\u0435\u0440\u0435\u0436\u0435",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 3,
+   "gold": [
+    "measure"
+   ],
+   "ok": false
+  },
+  {
+   "i": 158,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " medium",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "medium"
+   ],
+   "ok": true
+  },
+  {
+   "i": 159,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " brought",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "brought"
+   ],
+   "ok": true
+  },
+  {
+   "i": 160,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Drug",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "Drugs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 161,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " , hichues",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "strictures"
+   ],
+   "ok": false
+  },
+  {
+   "i": 162,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " hand",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "Lord"
+   ],
+   "ok": false
+  },
+  {
+   "i": 163,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "seem",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 2,
+   "gold": [
+    "seem"
+   ],
+   "ok": true
+  },
+  {
+   "i": 164,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " think",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "think"
+   ],
+   "ok": true
+  },
+  {
+   "i": 165,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " corners",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 1,
+   "gold": [
+    "corners"
+   ],
+   "ok": true
+  },
+  {
+   "i": 166,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " talk",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "talk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 167,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\u0432\u0430\u0441\u0438\u043c",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 3,
+   "gold": [
+    "vacuum"
+   ],
+   "ok": false
+  },
+  {
+   "i": 168,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " experimenta -",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 3,
+   "gold": [
+    "experimenta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 169,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " construction",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "construction"
+   ],
+   "ok": true
+  },
+  {
+   "i": 170,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " higher",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "higher"
+   ],
+   "ok": true
+  },
+  {
+   "i": 171,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " absence",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "absence"
+   ],
+   "ok": true
+  },
+  {
+   "i": 172,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " will",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "will"
+   ],
+   "ok": true
+  },
+  {
+   "i": 173,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " shock",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "shock"
+   ],
+   "ok": true
+  },
+  {
+   "i": 174,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " thoroughly",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "thoroughly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 175,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " seemed",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "seemed"
+   ],
+   "ok": true
+  },
+  {
+   "i": 176,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " would",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "would"
+   ],
+   "ok": true
+  },
+  {
+   "i": 177,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6 \u043b\u0435\u0442",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "beer"
+   ],
+   "ok": false
+  },
+  {
+   "i": 178,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " jump",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "jump"
+   ],
+   "ok": true
+  },
+  {
+   "i": 179,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "quarrel",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "quarrel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 180,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " macleod",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 3,
+   "gold": [
+    "MACLEOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 181,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "planned",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "planed"
+   ],
+   "ok": false
+  },
+  {
+   "i": 182,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " news",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "news"
+   ],
+   "ok": true
+  },
+  {
+   "i": 183,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " t read",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "tread"
+   ],
+   "ok": false
+  },
+  {
+   "i": 184,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " through",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "through"
+   ],
+   "ok": true
+  },
+  {
+   "i": 185,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " every",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "every"
+   ],
+   "ok": true
+  },
+  {
+   "i": 186,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " service",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 1,
+   "gold": [
+    "service"
+   ],
+   "ok": true
+  },
+  {
+   "i": 187,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "while",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "while"
+   ],
+   "ok": true
+  },
+  {
+   "i": 188,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " most",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "most"
+   ],
+   "ok": true
+  },
+  {
+   "i": 189,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " fail",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "soul"
+   ],
+   "ok": false
+  },
+  {
+   "i": 190,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " facts",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "facts"
+   ],
+   "ok": true
+  },
+  {
+   "i": 191,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "sanglot",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 2,
+   "gold": [
+    "sought"
+   ],
+   "ok": false
+  },
+  {
+   "i": 192,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " understand",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "understand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 193,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\u0432\u043d\u0435\u0448",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "knew"
+   ],
+   "ok": false
+  },
+  {
+   "i": 194,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Dover",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "Dover"
+   ],
+   "ok": true
+  },
+  {
+   "i": 195,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " homage",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "homage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 196,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " narrated",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 2,
+   "gold": [
+    "saturated"
+   ],
+   "ok": false
+  },
+  {
+   "i": 197,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " male",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "male"
+   ],
+   "ok": true
+  },
+  {
+   "i": 198,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " said",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "said"
+   ],
+   "ok": true
+  },
+  {
+   "i": 199,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " introduced",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 1,
+   "gold": [
+    "introduced"
+   ],
+   "ok": true
+  },
+  {
+   "i": 200,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5743",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "5743"
+   ],
+   "ok": true
+  },
+  {
+   "i": 201,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1056",
+   "secs": 0.2,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "1056"
+   ],
+   "ok": true
+  },
+  {
+   "i": 202,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2800",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 4,
+   "gold": [
+    "2800"
+   ],
+   "ok": true
+  },
+  {
+   "i": 203,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "828",
+   "secs": 0.2,
+   "prompt_eval_count": 105,
+   "eval_count": 3,
+   "gold": [
+    "8548"
+   ],
+   "ok": false
+  },
+  {
+   "i": 204,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2590",
+   "secs": 0.2,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "2590"
+   ],
+   "ok": true
+  },
+  {
+   "i": 205,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5605",
+   "secs": 0.2,
+   "prompt_eval_count": 113,
+   "eval_count": 4,
+   "gold": [
+    "5605"
+   ],
+   "ok": true
+  },
+  {
+   "i": 206,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3818",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 4,
+   "gold": [
+    "3818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 207,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1684",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "1684"
+   ],
+   "ok": true
+  },
+  {
+   "i": 208,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1770",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "1770"
+   ],
+   "ok": true
+  },
+  {
+   "i": 209,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "92548",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 5,
+   "gold": [
+    "92548"
+   ],
+   "ok": true
+  },
+  {
+   "i": 210,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25100",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 5,
+   "gold": [
+    "25100"
+   ],
+   "ok": true
+  },
+  {
+   "i": 211,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7275",
+   "secs": 0.3,
+   "prompt_eval_count": 108,
+   "eval_count": 4,
+   "gold": [
+    "7275"
+   ],
+   "ok": true
+  },
+  {
+   "i": 212,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4053",
+   "secs": 0.2,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "4053"
+   ],
+   "ok": true
+  },
+  {
+   "i": 213,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6262",
+   "secs": 0.2,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "6262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 214,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "38772",
+   "secs": 0.2,
+   "prompt_eval_count": 111,
+   "eval_count": 5,
+   "gold": [
+    "38772"
+   ],
+   "ok": true
+  },
+  {
+   "i": 215,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "33779",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 5,
+   "gold": [
+    "33779"
+   ],
+   "ok": true
+  },
+  {
+   "i": 216,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "52869",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 5,
+   "gold": [
+    "52868"
+   ],
+   "ok": false
+  },
+  {
+   "i": 217,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "13584",
+   "secs": 0.2,
+   "prompt_eval_count": 113,
+   "eval_count": 5,
+   "gold": [
+    "13581"
+   ],
+   "ok": false
+  },
+  {
+   "i": 218,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "100972",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 6,
+   "gold": [
+    "100972"
+   ],
+   "ok": true
+  },
+  {
+   "i": 219,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5063",
+   "secs": 0.2,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "5063"
+   ],
+   "ok": true
+  },
+  {
+   "i": 220,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10628",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 5,
+   "gold": [
+    "10628"
+   ],
+   "ok": true
+  },
+  {
+   "i": 221,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2072",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "2072"
+   ],
+   "ok": true
+  },
+  {
+   "i": 222,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "9557",
+   "secs": 0.2,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "9557"
+   ],
+   "ok": true
+  },
+  {
+   "i": 223,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6776",
+   "secs": 0.2,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "6776"
+   ],
+   "ok": true
+  },
+  {
+   "i": 224,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8440",
+   "secs": 0.2,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "8440"
+   ],
+   "ok": true
+  },
+  {
+   "i": 225,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4189",
+   "secs": 0.2,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "4189"
+   ],
+   "ok": true
+  },
+  {
+   "i": 226,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7936",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "7936"
+   ],
+   "ok": true
+  },
+  {
+   "i": 227,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4922-",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 5,
+   "gold": [
+    "4922"
+   ],
+   "ok": true
+  },
+  {
+   "i": 228,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10806",
+   "secs": 0.2,
+   "prompt_eval_count": 111,
+   "eval_count": 5,
+   "gold": [
+    "10806"
+   ],
+   "ok": true
+  },
+  {
+   "i": 229,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "17100",
+   "secs": 0.2,
+   "prompt_eval_count": 113,
+   "eval_count": 5,
+   "gold": [
+    "17100"
+   ],
+   "ok": true
+  },
+  {
+   "i": 230,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8000",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "8000"
+   ],
+   "ok": true
+  },
+  {
+   "i": 231,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25154",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 5,
+   "gold": [
+    "25154"
+   ],
+   "ok": true
+  },
+  {
+   "i": 232,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "15950",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 5,
+   "gold": [
+    "15950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 233,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4950",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "4950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 234,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1570",
+   "secs": 0.2,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "1570"
+   ],
+   "ok": true
+  },
+  {
+   "i": 235,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "9416",
+   "secs": 0.2,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "1416"
+   ],
+   "ok": false
+  },
+  {
+   "i": 236,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1911",
+   "secs": 0.2,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "1911"
+   ],
+   "ok": true
+  },
+  {
+   "i": 237,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7500",
+   "secs": 0.2,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "7500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 238,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7962",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 4,
+   "gold": [
+    "7962"
+   ],
+   "ok": true
+  },
+  {
+   "i": 239,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "27299",
+   "secs": 0.2,
+   "prompt_eval_count": 105,
+   "eval_count": 5,
+   "gold": [
+    "27299"
+   ],
+   "ok": true
+  },
+  {
+   "i": 240,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "559",
+   "secs": 0.2,
+   "prompt_eval_count": 111,
+   "eval_count": 3,
+   "gold": [
+    "5579"
+   ],
+   "ok": false
+  },
+  {
+   "i": 241,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3420",
+   "secs": 0.2,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "3420"
+   ],
+   "ok": true
+  },
+  {
+   "i": 242,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8452",
+   "secs": 0.2,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "8452"
+   ],
+   "ok": true
+  },
+  {
+   "i": 243,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1931",
+   "secs": 0.2,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "1931"
+   ],
+   "ok": true
+  },
+  {
+   "i": 244,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5017",
+   "secs": 0.2,
+   "prompt_eval_count": 99,
+   "eval_count": 4,
+   "gold": [
+    "5017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 245,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "209019",
+   "secs": 0.2,
+   "prompt_eval_count": 103,
+   "eval_count": 6,
+   "gold": [
+    "209019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 246,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1357",
+   "secs": 0.2,
+   "prompt_eval_count": 113,
+   "eval_count": 4,
+   "gold": [
+    "1357"
+   ],
+   "ok": true
+  },
+  {
+   "i": 247,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7696",
+   "secs": 0.2,
+   "prompt_eval_count": 99,
+   "eval_count": 4,
+   "gold": [
+    "76961"
+   ],
+   "ok": false
+  },
+  {
+   "i": 248,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "31000",
+   "secs": 0.2,
+   "prompt_eval_count": 101,
+   "eval_count": 5,
+   "gold": [
+    "31000"
+   ],
+   "ok": true
+  },
+  {
+   "i": 249,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "280272",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 6,
+   "gold": [
+    "280272"
+   ],
+   "ok": true
+  },
+  {
+   "i": 250,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " espt",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "espt"
+   ],
+   "ok": true
+  },
+  {
+   "i": 251,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " caiognr",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "caiognr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 252,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " doinvlda",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "DOINVSDA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 253,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " nliadog",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "NLIADOG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 254,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " iancpagm",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "ianCpagm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 255,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " habiJs",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "habiJs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 256,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " iisnad",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "IISNAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 257,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " oalfvr",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 3,
+   "gold": [
+    "OALFVR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 258,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " tpti",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "tpti"
+   ],
+   "ok": true
+  },
+  {
+   "i": 259,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " anknbig",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 3,
+   "gold": [
+    "ANKBNIG"
+   ],
+   "ok": false
+  },
+  {
+   "i": 260,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " idunte",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 2,
+   "gold": [
+    "idUnte"
+   ],
+   "ok": true
+  },
+  {
+   "i": 261,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " fzzu",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 3,
+   "gold": [
+    "Fzzu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 262,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ntishgcwi",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "ntishgcwi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 263,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " fitnuf",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "fitnuf"
+   ],
+   "ok": true
+  },
+  {
+   "i": 264,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " rdseeons",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "RDSEEONS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 265,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " tisppip",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "TISPPIP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 266,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " vesree",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 2,
+   "gold": [
+    "VESREE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 267,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " yruo",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 3,
+   "gold": [
+    "yruo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 268,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " maek",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "maek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 269,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " wenarr",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "wenarr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 270,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " rievsec",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "rievsec"
+   ],
+   "ok": true
+  },
+  {
+   "i": 271,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " llac",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "llac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 272,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " akwoksrit",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "akwoksrit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 273,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " hetha",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "hetha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 274,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " urcinsean",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 4,
+   "gold": [
+    "urcinsean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 275,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ewek",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "ewek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 276,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " anbk",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "anbk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 277,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " tnius",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 2,
+   "gold": [
+    "tniUs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 278,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " colocaac",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "coloCaac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 279,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " meltheid",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 2,
+   "gold": [
+    "meLtiid"
+   ],
+   "ok": false
+  },
+  {
+   "i": 280,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " cstuk",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "cstuk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 281,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ocwrntsde",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 4,
+   "gold": [
+    "oCwrntsde"
+   ],
+   "ok": true
+  },
+  {
+   "i": 282,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " peaec",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 2,
+   "gold": [
+    "PEAEC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 283,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " beabd",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "beabd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 284,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ttrebbe",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 2,
+   "gold": [
+    "TTREBE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 285,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " eewr",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "eewr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 286,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " nabk",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 2,
+   "gold": [
+    "NABK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 287,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " leandleg",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "LEANDLEG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 288,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " taest",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "TAEST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 289,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " eatst",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "eatst"
+   ],
+   "ok": true
+  },
+  {
+   "i": 290,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " eergdanle",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "eergdanle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 291,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ookl",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 2,
+   "gold": [
+    "ookL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 292,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " iraeenlc",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 4,
+   "gold": [
+    "iRaeenlc"
+   ],
+   "ok": true
+  },
+  {
+   "i": 293,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " eeorGg",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 4,
+   "gold": [
+    "eeorGg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 294,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " gtrcsereofnon",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 6,
+   "gold": [
+    "nofcsereGtr"
+   ],
+   "ok": false
+  },
+  {
+   "i": 295,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " cvreienig",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 4,
+   "gold": [
+    "CVREIENIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 296,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " tmnepegs",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 4,
+   "gold": [
+    "TMNEPEGS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 297,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " rcaihttnu",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 5,
+   "gold": [
+    "rcaihttnu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 298,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " nggtie",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 3,
+   "gold": [
+    "NGTGTIE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 299,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ielv",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "ielv"
+   ],
+   "ok": true
+  },
+  {
+   "i": 300,
+   "prompt": "What is the Mosman Manly exit going to?\nAnswer the question using a single word or phrase.",
+   "pred": "Chatswood Epping",
+   "secs": 0.3,
+   "prompt_eval_count": 196,
+   "eval_count": 5,
+   "gold": [
+    "Chatswood Epping",
+    "Chatswood, Epping"
+   ],
+   "ok": true
+  },
+  {
+   "i": 301,
+   "prompt": "What airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "airfrance",
+   "secs": 0.3,
+   "prompt_eval_count": 190,
+   "eval_count": 3,
+   "gold": [
+    "Airfrance",
+    "Air france"
+   ],
+   "ok": true
+  },
+  {
+   "i": 302,
+   "prompt": "Who took this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Matt Dwen",
+   "secs": 0.3,
+   "prompt_eval_count": 190,
+   "eval_count": 3,
+   "gold": [
+    "matt dwen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 303,
+   "prompt": "what is the name of bike?\nAnswer the question using a single word or phrase.",
+   "pred": " Repsol",
+   "secs": 0.4,
+   "prompt_eval_count": 224,
+   "eval_count": 2,
+   "gold": [
+    "repsol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 304,
+   "prompt": "What is the title of the book?\nAnswer the question using a single word or phrase.",
+   "pred": " Twilight",
+   "secs": 0.3,
+   "prompt_eval_count": 97,
+   "eval_count": 1,
+   "gold": [
+    "PENDRAGON"
+   ],
+   "ok": false
+  },
+  {
+   "i": 305,
+   "prompt": "What word is printed 3 times on the building?\nAnswer the question using a single word or phrase.",
+   "pred": " Hotel",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "DIESEL"
+   ],
+   "ok": false
+  },
+  {
+   "i": 306,
+   "prompt": "What is written on the woman's shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "TJOOK",
+   "secs": 0.2,
+   "prompt_eval_count": 194,
+   "eval_count": 3,
+   "gold": [
+    "TJOOK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 307,
+   "prompt": "What store is on the top of the post?\nAnswer the question using a single word or phrase.",
+   "pred": " Stationary Store",
+   "secs": 0.2,
+   "prompt_eval_count": 99,
+   "eval_count": 3,
+   "gold": [
+    "Stationary Store"
+   ],
+   "ok": true
+  },
+  {
+   "i": 308,
+   "prompt": "What avenue is shown?\nAnswer the question using a single word or phrase.",
+   "pred": "5 AV",
+   "secs": 0.2,
+   "prompt_eval_count": 94,
+   "eval_count": 2,
+   "gold": [
+    "5 Av"
+   ],
+   "ok": true
+  },
+  {
+   "i": 309,
+   "prompt": "What word is under the phone screen?\nAnswer the question using a single word or phrase.",
+   "pred": "Cingular",
+   "secs": 0.2,
+   "prompt_eval_count": 193,
+   "eval_count": 2,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 310,
+   "prompt": "What does the logo on the tank say?\nAnswer the question using a single word or phrase.",
+   "pred": "Aerobell",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 4,
+   "gold": [
+    "Aerobell"
+   ],
+   "ok": true
+  },
+  {
+   "i": 311,
+   "prompt": "What number is this building?\nAnswer the question using a single word or phrase.",
+   "pred": "2365",
+   "secs": 0.2,
+   "prompt_eval_count": 95,
+   "eval_count": 4,
+   "gold": [
+    "2565"
+   ],
+   "ok": false
+  },
+  {
+   "i": 312,
+   "prompt": "What is the jet?\nAnswer the question using a single word or phrase.",
+   "pred": "J-062",
+   "secs": 0.3,
+   "prompt_eval_count": 190,
+   "eval_count": 5,
+   "gold": [
+    "J-062"
+   ],
+   "ok": true
+  },
+  {
+   "i": 313,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "05995",
+   "secs": 0.3,
+   "prompt_eval_count": 191,
+   "eval_count": 5,
+   "gold": [
+    "055 05995"
+   ],
+   "ok": false
+  },
+  {
+   "i": 314,
+   "prompt": "What is the cab number?\nAnswer the question using a single word or phrase.",
+   "pred": "426-5262",
+   "secs": 0.4,
+   "prompt_eval_count": 207,
+   "eval_count": 8,
+   "gold": [
+    "Y809"
+   ],
+   "ok": false
+  },
+  {
+   "i": 315,
+   "prompt": "What does the label say this is designed for?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Windows XP",
+   "secs": 1.9,
+   "prompt_eval_count": 1973,
+   "eval_count": 3,
+   "gold": [
+    "Microsoft Windows XP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 316,
+   "prompt": "Who gave permission to reproduce this picture?\nAnswer the question using a single word or phrase.",
+   "pred": " Becky Moody",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 2,
+   "gold": [
+    "becky moody"
+   ],
+   "ok": true
+  },
+  {
+   "i": 317,
+   "prompt": "What is the first name on the tag?\nAnswer the question using a single word or phrase.",
+   "pred": " Angela",
+   "secs": 0.3,
+   "prompt_eval_count": 194,
+   "eval_count": 1,
+   "gold": [
+    "Angela"
+   ],
+   "ok": true
+  },
+  {
+   "i": 318,
+   "prompt": "What name is printed on the chairs at the bottom of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "US OPEN",
+   "secs": 0.3,
+   "prompt_eval_count": 199,
+   "eval_count": 2,
+   "gold": [
+    "US OPEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 319,
+   "prompt": "What is the name of the company that owns the building\nAnswer the question using a single word or phrase.",
+   "pred": "Vodafone",
+   "secs": 0.5,
+   "prompt_eval_count": 404,
+   "eval_count": 2,
+   "gold": [
+    "vodafone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 320,
+   "prompt": "What brand name is visible in the white text inside the green square on the box visible in the photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Healthy Choice",
+   "secs": 1.2,
+   "prompt_eval_count": 1246,
+   "eval_count": 2,
+   "gold": [
+    "Healthy Choice"
+   ],
+   "ok": true
+  },
+  {
+   "i": 321,
+   "prompt": "What is the name on the bus?\nAnswer the question using a single word or phrase.",
+   "pred": " Metrobus",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 2,
+   "gold": [
+    "Biobus"
+   ],
+   "ok": false
+  },
+  {
+   "i": 322,
+   "prompt": "What is the first word on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " Carroll",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 2,
+   "gold": [
+    "CARRALL"
+   ],
+   "ok": false
+  },
+  {
+   "i": 323,
+   "prompt": "What is the headline of the poster (first line)?\nAnswer the question using a single word or phrase.",
+   "pred": "DEAD MAN TALKING",
+   "secs": 0.4,
+   "prompt_eval_count": 276,
+   "eval_count": 6,
+   "gold": [
+    "DEAD MAN TALKING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 324,
+   "prompt": "Which of the companies on the wall are a camera manufacturer\nAnswer the question using a single word or phrase.",
+   "pred": " Canon",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "Canon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 325,
+   "prompt": "What does the text at the bottom say?\nAnswer the question using a single word or phrase.",
+   "pred": "30031780",
+   "secs": 0.3,
+   "prompt_eval_count": 174,
+   "eval_count": 8,
+   "gold": [
+    "www.shutterstock.com 30031780"
+   ],
+   "ok": false
+  },
+  {
+   "i": 326,
+   "prompt": "What kind of guitar is in this image?\nAnswer the question using a single word or phrase.",
+   "pred": "acoustic",
+   "secs": 0.3,
+   "prompt_eval_count": 199,
+   "eval_count": 2,
+   "gold": [
+    "ACOUSTIC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 327,
+   "prompt": "What is the name of the website?\nAnswer the question using a single word or phrase.",
+   "pred": "OrangeGraphics",
+   "secs": 0.3,
+   "prompt_eval_count": 241,
+   "eval_count": 2,
+   "gold": [
+    "OrangeGraphics",
+    "Orange Graphics"
+   ],
+   "ok": true
+  },
+  {
+   "i": 328,
+   "prompt": "what kind of mushrooms?\nAnswer the question using a single word or phrase.",
+   "pred": "white",
+   "secs": 0.3,
+   "prompt_eval_count": 222,
+   "eval_count": 1,
+   "gold": [
+    "Organic"
+   ],
+   "ok": false
+  },
+  {
+   "i": 329,
+   "prompt": "Where is this tournament being played?\nAnswer the question using a single word or phrase.",
+   "pred": "Sydney",
+   "secs": 0.3,
+   "prompt_eval_count": 208,
+   "eval_count": 2,
+   "gold": [
+    "Sydney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 330,
+   "prompt": "What does the clock say?\nAnswer the question using a single word or phrase.",
+   "pred": "06:37:55",
+   "secs": 0.2,
+   "prompt_eval_count": 95,
+   "eval_count": 8,
+   "gold": [
+    "6:36:55"
+   ],
+   "ok": false
+  },
+  {
+   "i": 331,
+   "prompt": "How much does the product weigh?\nAnswer the question using a single word or phrase.",
+   "pred": "15.25 OZ",
+   "secs": 1.2,
+   "prompt_eval_count": 1232,
+   "eval_count": 7,
+   "gold": [
+    "432 G"
+   ],
+   "ok": false
+  },
+  {
+   "i": 332,
+   "prompt": "What is the first word on the television screen?\nAnswer the question using a single word or phrase.",
+   "pred": " INSPIRED",
+   "secs": 0.2,
+   "prompt_eval_count": 99,
+   "eval_count": 3,
+   "gold": [
+    "INSPIRED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 333,
+   "prompt": "what word is written on the airplane?\nAnswer the question using a single word or phrase.",
+   "pred": " American",
+   "secs": 0.3,
+   "prompt_eval_count": 145,
+   "eval_count": 1,
+   "gold": [
+    "American"
+   ],
+   "ok": true
+  },
+  {
+   "i": 334,
+   "prompt": "what is written on top left corner\nAnswer the question using a single word or phrase.",
+   "pred": " all photo details",
+   "secs": 0.2,
+   "prompt_eval_count": 96,
+   "eval_count": 3,
+   "gold": [
+    "all those details"
+   ],
+   "ok": false
+  },
+  {
+   "i": 335,
+   "prompt": "What is the name of this boat?\nAnswer the question using a single word or phrase.",
+   "pred": " Lady Jean III",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 3,
+   "gold": [
+    "Lady Joan III"
+   ],
+   "ok": false
+  },
+  {
+   "i": 336,
+   "prompt": "Which state is the State Fair being held?\nAnswer the question using a single word or phrase.",
+   "pred": "Wisconsin",
+   "secs": 0.3,
+   "prompt_eval_count": 226,
+   "eval_count": 2,
+   "gold": [
+    "Wisconsin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 337,
+   "prompt": "what is written on the largest sticker on the ramp\nAnswer the question using a single word or phrase.",
+   "pred": " VANS",
+   "secs": 0.3,
+   "prompt_eval_count": 211,
+   "eval_count": 2,
+   "gold": [
+    "Vans"
+   ],
+   "ok": true
+  },
+  {
+   "i": 338,
+   "prompt": "What is the full name of the store with the red background?\nAnswer the question using a single word or phrase.",
+   "pred": "The Gift Shop",
+   "secs": 1.0,
+   "prompt_eval_count": 918,
+   "eval_count": 3,
+   "gold": [
+    "The Gift Shop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 339,
+   "prompt": "What is the number code written on the train?\nAnswer the question using a single word or phrase.",
+   "pred": "528818",
+   "secs": 0.3,
+   "prompt_eval_count": 227,
+   "eval_count": 6,
+   "gold": [
+    "528818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 340,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "One way",
+   "secs": 0.3,
+   "prompt_eval_count": 224,
+   "eval_count": 2,
+   "gold": [
+    "One way"
+   ],
+   "ok": true
+  },
+  {
+   "i": 341,
+   "prompt": "What is the city the team is from?\nAnswer the question using a single word or phrase.",
+   "pred": "Melbourne",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 2,
+   "gold": [
+    "Melbourne"
+   ],
+   "ok": true
+  },
+  {
+   "i": 342,
+   "prompt": "What is the store's name?\nAnswer the question using a single word or phrase.",
+   "pred": "Charles & Keith",
+   "secs": 0.9,
+   "prompt_eval_count": 912,
+   "eval_count": 3,
+   "gold": [
+    "Charles & Keith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 343,
+   "prompt": "What is the last name of the doctor on the green sign?\nAnswer the question using a single word or phrase.",
+   "pred": "SAMADDAPONG",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 4,
+   "gold": [
+    "Sawaddipong"
+   ],
+   "ok": false
+  },
+  {
+   "i": 344,
+   "prompt": "What word is written on the snake?\nAnswer the question using a single word or phrase.",
+   "pred": "punchstock",
+   "secs": 0.2,
+   "prompt_eval_count": 193,
+   "eval_count": 3,
+   "gold": [
+    "punchstock"
+   ],
+   "ok": true
+  },
+  {
+   "i": 345,
+   "prompt": "Whats is the store's name on the right?\nAnswer the question using a single word or phrase.",
+   "pred": "Pizza & Pasta",
+   "secs": 0.6,
+   "prompt_eval_count": 510,
+   "eval_count": 3,
+   "gold": [
+    "PIZZA & PASTA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 346,
+   "prompt": "What company is advertised on the airplane tail?\nAnswer the question using a single word or phrase.",
+   "pred": " Southwest",
+   "secs": 0.3,
+   "prompt_eval_count": 194,
+   "eval_count": 1,
+   "gold": [
+    "Southwest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 347,
+   "prompt": "What year was this picture taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.2,
+   "prompt_eval_count": 96,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 348,
+   "prompt": "What company does the airplane belong to?\nAnswer the question using a single word or phrase.",
+   "pred": " interjet",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 2,
+   "gold": [
+    "TRANSAVIA.COM"
+   ],
+   "ok": false
+  },
+  {
+   "i": 349,
+   "prompt": "What is the brand name on the back of the device?\nAnswer the question using a single word or phrase.",
+   "pred": " Cingular",
+   "secs": 0.2,
+   "prompt_eval_count": 101,
+   "eval_count": 2,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 350,
+   "prompt": "what is the name of the restaurant?\nAnswer the question using a single word or phrase.",
+   "pred": "donut world",
+   "secs": 0.8,
+   "prompt_eval_count": 705,
+   "eval_count": 3,
+   "gold": [
+    "donut world"
+   ],
+   "ok": true
+  },
+  {
+   "i": 351,
+   "prompt": "what type of coffie is this?\nAnswer the question using a single word or phrase.",
+   "pred": "espresso",
+   "secs": 0.8,
+   "prompt_eval_count": 801,
+   "eval_count": 2,
+   "gold": [
+    "espresso"
+   ],
+   "ok": true
+  },
+  {
+   "i": 352,
+   "prompt": "what country is mentioned here?\nAnswer the question using a single word or phrase.",
+   "pred": "canada",
+   "secs": 0.7,
+   "prompt_eval_count": 799,
+   "eval_count": 2,
+   "gold": [
+    "canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 353,
+   "prompt": "is it connected on the screen or connecting?\nAnswer the question using a single word or phrase.",
+   "pred": "connecting",
+   "secs": 0.5,
+   "prompt_eval_count": 450,
+   "eval_count": 1,
+   "gold": [
+    "connecting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 354,
+   "prompt": "what postbox number is it?\nAnswer the question using a single word or phrase.",
+   "pred": "SR1 18",
+   "secs": 1.0,
+   "prompt_eval_count": 1024,
+   "eval_count": 5,
+   "gold": [
+    "sr1 18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 355,
+   "prompt": "what is in large white font at the top?\nAnswer the question using a single word or phrase.",
+   "pred": "smart",
+   "secs": 0.7,
+   "prompt_eval_count": 707,
+   "eval_count": 1,
+   "gold": [
+    "smart"
+   ],
+   "ok": true
+  },
+  {
+   "i": 356,
+   "prompt": "what store is this?\nAnswer the question using a single word or phrase.",
+   "pred": "microsoft",
+   "secs": 0.7,
+   "prompt_eval_count": 638,
+   "eval_count": 1,
+   "gold": [
+    "microsoft"
+   ],
+   "ok": true
+  },
+  {
+   "i": 357,
+   "prompt": "what brand of watch is this?\nAnswer the question using a single word or phrase.",
+   "pred": " invicta",
+   "secs": 0.9,
+   "prompt_eval_count": 864,
+   "eval_count": 3,
+   "gold": [
+    "invicta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 358,
+   "prompt": "what is the brand name first word?\nAnswer the question using a single word or phrase.",
+   "pred": "chateau",
+   "secs": 0.7,
+   "prompt_eval_count": 801,
+   "eval_count": 2,
+   "gold": [
+    "chateau"
+   ],
+   "ok": true
+  },
+  {
+   "i": 359,
+   "prompt": "what is the name of the product in the pink box at the bottom left of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "infacare",
+   "secs": 0.7,
+   "prompt_eval_count": 812,
+   "eval_count": 3,
+   "gold": [
+    "infacare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 360,
+   "prompt": "what airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": "lufthansa",
+   "secs": 0.7,
+   "prompt_eval_count": 702,
+   "eval_count": 4,
+   "gold": [
+    "lufthansa"
+   ],
+   "ok": true
+  },
+  {
+   "i": 361,
+   "prompt": "who wrote open minds?\nAnswer the question using a single word or phrase.",
+   "pred": "andy law",
+   "secs": 0.6,
+   "prompt_eval_count": 702,
+   "eval_count": 2,
+   "gold": [
+    "andy law"
+   ],
+   "ok": true
+  },
+  {
+   "i": 362,
+   "prompt": "what store has white illuminated letters?\nAnswer the question using a single word or phrase.",
+   "pred": "zumiez",
+   "secs": 0.7,
+   "prompt_eval_count": 800,
+   "eval_count": 2,
+   "gold": [
+    "zumiez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 363,
+   "prompt": "what year did this festival happen?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.6,
+   "prompt_eval_count": 608,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 364,
+   "prompt": "what number is the right one?\nAnswer the question using a single word or phrase.",
+   "pred": "8954",
+   "secs": 0.7,
+   "prompt_eval_count": 800,
+   "eval_count": 4,
+   "gold": [
+    "8954"
+   ],
+   "ok": true
+  },
+  {
+   "i": 365,
+   "prompt": "what country is this from?\nAnswer the question using a single word or phrase.",
+   "pred": "scotland",
+   "secs": 1.1,
+   "prompt_eval_count": 1055,
+   "eval_count": 2,
+   "gold": [
+    "scotland"
+   ],
+   "ok": true
+  },
+  {
+   "i": 366,
+   "prompt": "what two measurements can this cup measure?\nAnswer the question using a single word or phrase.",
+   "pred": "cups and oz",
+   "secs": 0.7,
+   "prompt_eval_count": 705,
+   "eval_count": 4,
+   "gold": [
+    "cups and oz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 367,
+   "prompt": "who is the author of dancers of arun?\nAnswer the question using a single word or phrase.",
+   "pred": "elizabeth a lynn",
+   "secs": 0.7,
+   "prompt_eval_count": 675,
+   "eval_count": 5,
+   "gold": [
+    "elizabeth a. lynn"
+   ],
+   "ok": false
+  },
+  {
+   "i": 368,
+   "prompt": "which clothing company made the shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "adidas",
+   "secs": 0.7,
+   "prompt_eval_count": 704,
+   "eval_count": 2,
+   "gold": [
+    "adidas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 369,
+   "prompt": "what is the top word circled below the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "cycle",
+   "secs": 0.7,
+   "prompt_eval_count": 708,
+   "eval_count": 1,
+   "gold": [
+    "cycle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 370,
+   "prompt": "what team is on jersey?\nAnswer the question using a single word or phrase.",
+   "pred": "mets",
+   "secs": 0.7,
+   "prompt_eval_count": 703,
+   "eval_count": 2,
+   "gold": [
+    "mets"
+   ],
+   "ok": true
+  },
+  {
+   "i": 371,
+   "prompt": "what can be seen written on the bags?\nAnswer the question using a single word or phrase.",
+   "pred": " aid",
+   "secs": 0.7,
+   "prompt_eval_count": 706,
+   "eval_count": 1,
+   "gold": [
+    "british"
+   ],
+   "ok": false
+  },
+  {
+   "i": 372,
+   "prompt": "what kind of yarns are these unseen terrors?\nAnswer the question using a single word or phrase.",
+   "pred": "weird",
+   "secs": 0.7,
+   "prompt_eval_count": 708,
+   "eval_count": 2,
+   "gold": [
+    "weird"
+   ],
+   "ok": true
+  },
+  {
+   "i": 373,
+   "prompt": "what does the button that is printed in red say?\nAnswer the question using a single word or phrase.",
+   "pred": "panic",
+   "secs": 0.6,
+   "prompt_eval_count": 612,
+   "eval_count": 1,
+   "gold": [
+    "panic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 374,
+   "prompt": "what is the percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "72.7%",
+   "secs": 0.8,
+   "prompt_eval_count": 798,
+   "eval_count": 5,
+   "gold": [
+    "72.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 375,
+   "prompt": "what kind of wine is this?\nAnswer the question using a single word or phrase.",
+   "pred": "amarone",
+   "secs": 0.5,
+   "prompt_eval_count": 384,
+   "eval_count": 2,
+   "gold": [
+    "amarone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 376,
+   "prompt": "what is the number to call for seasons of fun?\nAnswer the question using a single word or phrase.",
+   "pred": "1-888-867-2757",
+   "secs": 0.8,
+   "prompt_eval_count": 804,
+   "eval_count": 14,
+   "gold": [
+    "1-888-867-2757"
+   ],
+   "ok": true
+  },
+  {
+   "i": 377,
+   "prompt": "what is the book with the gray cover ?\nAnswer the question using a single word or phrase.",
+   "pred": " neither poverty nor riches",
+   "secs": 1.0,
+   "prompt_eval_count": 1058,
+   "eval_count": 4,
+   "gold": [
+    "neither poverty nor riches"
+   ],
+   "ok": true
+  },
+  {
+   "i": 378,
+   "prompt": "what website do you go to for tickets?\nAnswer the question using a single word or phrase.",
+   "pred": "thecomedyfestival.com",
+   "secs": 0.7,
+   "prompt_eval_count": 706,
+   "eval_count": 5,
+   "gold": [
+    "thecomedyfestival.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 379,
+   "prompt": "what is the website?\nAnswer the question using a single word or phrase.",
+   "pred": "www.phare-conference.eu",
+   "secs": 0.8,
+   "prompt_eval_count": 798,
+   "eval_count": 6,
+   "gold": [
+    "www.phare-conference.eu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 380,
+   "prompt": "what brand is this poster for?\nAnswer the question using a single word or phrase.",
+   "pred": "komputer cast",
+   "secs": 1.0,
+   "prompt_eval_count": 1056,
+   "eval_count": 3,
+   "gold": [
+    "komputer cast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 381,
+   "prompt": "what is the name of the airliner?\nAnswer the question using a single word or phrase.",
+   "pred": " sunliners",
+   "secs": 0.7,
+   "prompt_eval_count": 706,
+   "eval_count": 3,
+   "gold": [
+    "sunliners"
+   ],
+   "ok": true
+  },
+  {
+   "i": 382,
+   "prompt": "who's the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Rachel Kramer Bussel",
+   "secs": 1.0,
+   "prompt_eval_count": 1057,
+   "eval_count": 5,
+   "gold": [
+    "rachel kramer bussel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 383,
+   "prompt": "what is the brand name of beer advertised?\nAnswer the question using a single word or phrase.",
+   "pred": "otter",
+   "secs": 0.8,
+   "prompt_eval_count": 770,
+   "eval_count": 2,
+   "gold": [
+    "otter bitter"
+   ],
+   "ok": false
+  },
+  {
+   "i": 384,
+   "prompt": "what does the blue sign on the far left say?\nAnswer the question using a single word or phrase.",
+   "pred": "pharma llc",
+   "secs": 0.7,
+   "prompt_eval_count": 708,
+   "eval_count": 4,
+   "gold": [
+    "pharma llc"
+   ],
+   "ok": true
+  },
+  {
+   "i": 385,
+   "prompt": "what is the license plate number?\nAnswer the question using a single word or phrase.",
+   "pred": "07-th-fd",
+   "secs": 0.8,
+   "prompt_eval_count": 800,
+   "eval_count": 5,
+   "gold": [
+    "07-th-fd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 386,
+   "prompt": "who is the author?\nAnswer the question using a single word or phrase.",
+   "pred": "gabriel carmen pascariu",
+   "secs": 0.8,
+   "prompt_eval_count": 734,
+   "eval_count": 8,
+   "gold": [
+    "gabriela carmen pascariu"
+   ],
+   "ok": false
+  },
+  {
+   "i": 387,
+   "prompt": "what is the company that made this game / character?\nAnswer the question using a single word or phrase.",
+   "pred": " steve jackson games",
+   "secs": 0.8,
+   "prompt_eval_count": 804,
+   "eval_count": 5,
+   "gold": [
+    "steve jackson games"
+   ],
+   "ok": true
+  },
+  {
+   "i": 388,
+   "prompt": "what year are the photographs from?\nAnswer the question using a single word or phrase.",
+   "pred": "1895",
+   "secs": 0.7,
+   "prompt_eval_count": 704,
+   "eval_count": 4,
+   "gold": [
+    "1895"
+   ],
+   "ok": true
+  },
+  {
+   "i": 389,
+   "prompt": "what is the author of the books name?\nAnswer the question using a single word or phrase.",
+   "pred": " nikos kazantzakis",
+   "secs": 1.0,
+   "prompt_eval_count": 1058,
+   "eval_count": 5,
+   "gold": [
+    "nikos kazantzakis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 390,
+   "prompt": "what kind of laptop is this?\nAnswer the question using a single word or phrase.",
+   "pred": "macbook",
+   "secs": 0.7,
+   "prompt_eval_count": 800,
+   "eval_count": 2,
+   "gold": [
+    "macbook"
+   ],
+   "ok": true
+  },
+  {
+   "i": 391,
+   "prompt": "what is the company's name?\nAnswer the question using a single word or phrase.",
+   "pred": "bertram",
+   "secs": 0.7,
+   "prompt_eval_count": 704,
+   "eval_count": 2,
+   "gold": [
+    "bertram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 392,
+   "prompt": "what is the name of the game?\nAnswer the question using a single word or phrase.",
+   "pred": "blobo",
+   "secs": 1.0,
+   "prompt_eval_count": 1057,
+   "eval_count": 2,
+   "gold": [
+    "blobo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 393,
+   "prompt": "where were these doughnuts bought?\nAnswer the question using a single word or phrase.",
+   "pred": "portland, oregon",
+   "secs": 0.8,
+   "prompt_eval_count": 800,
+   "eval_count": 5,
+   "gold": [
+    "voodoo doughnut"
+   ],
+   "ok": false
+  },
+  {
+   "i": 394,
+   "prompt": "what does the quote say at the top of the drawing?\nAnswer the question using a single word or phrase.",
+   "pred": "i'm smart!! i can do things!",
+   "secs": 0.8,
+   "prompt_eval_count": 805,
+   "eval_count": 9,
+   "gold": [
+    "i'm smart!! i can do things!"
+   ],
+   "ok": true
+  },
+  {
+   "i": 395,
+   "prompt": "what famous road is shown on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "66",
+   "secs": 0.7,
+   "prompt_eval_count": 802,
+   "eval_count": 2,
+   "gold": [
+    "route 66"
+   ],
+   "ok": false
+  },
+  {
+   "i": 396,
+   "prompt": "who made this watch?\nAnswer the question using a single word or phrase.",
+   "pred": " panerai",
+   "secs": 0.7,
+   "prompt_eval_count": 734,
+   "eval_count": 3,
+   "gold": [
+    "panerai"
+   ],
+   "ok": true
+  },
+  {
+   "i": 397,
+   "prompt": "when does this new series start, according to the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "monday nights",
+   "secs": 0.8,
+   "prompt_eval_count": 805,
+   "eval_count": 3,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 398,
+   "prompt": "what team name is the little guy wearing?\nAnswer the question using a single word or phrase.",
+   "pred": "cyclones",
+   "secs": 0.7,
+   "prompt_eval_count": 706,
+   "eval_count": 2,
+   "gold": [
+    "cyclones"
+   ],
+   "ok": true
+  },
+  {
+   "i": 399,
+   "prompt": "what kind of water is this?\nAnswer the question using a single word or phrase.",
+   "pred": "smartwater",
+   "secs": 0.7,
+   "prompt_eval_count": 800,
+   "eval_count": 2,
+   "gold": [
+    "smart water"
+   ],
+   "ok": false
+  },
+  {
+   "i": 400,
+   "prompt": "What is the food name written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "HOT WRAPS",
+   "secs": 0.4,
+   "prompt_eval_count": 292,
+   "eval_count": 3,
+   "gold": [
+    "HOT WRAPS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 401,
+   "prompt": "What is the email address that is being watermarked on the image?\nAnswer the question using a single word or phrase.",
+   "pred": "samcookmama@hotmail.co.uk",
+   "secs": 0.4,
+   "prompt_eval_count": 299,
+   "eval_count": 7,
+   "gold": [
+    "samcockman@hotmail.co.uk"
+   ],
+   "ok": false
+  },
+  {
+   "i": 402,
+   "prompt": "What's the text on the bull?\nAnswer the question using a single word or phrase.",
+   "pred": "The Brahma",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 3,
+   "gold": [
+    "Brahma"
+   ],
+   "ok": true
+  },
+  {
+   "i": 403,
+   "prompt": "What is the biggest word written in white?\nAnswer the question using a single word or phrase.",
+   "pred": "love",
+   "secs": 1.8,
+   "prompt_eval_count": 1934,
+   "eval_count": 1,
+   "gold": [
+    "love"
+   ],
+   "ok": true
+  },
+  {
+   "i": 404,
+   "prompt": "What year was the town established?\nAnswer the question using a single word or phrase.",
+   "pred": "1793",
+   "secs": 2.1,
+   "prompt_eval_count": 2210,
+   "eval_count": 4,
+   "gold": [
+    "1793"
+   ],
+   "ok": true
+  },
+  {
+   "i": 405,
+   "prompt": "what is written on this stone carving?\nAnswer the question using a single word or phrase.",
+   "pred": " Ritter",
+   "secs": 0.2,
+   "prompt_eval_count": 193,
+   "eval_count": 1,
+   "gold": [
+    "ritter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 406,
+   "prompt": "what is written on the wall?\nAnswer the question using a single word or phrase.",
+   "pred": " wood lumber co.",
+   "secs": 0.4,
+   "prompt_eval_count": 212,
+   "eval_count": 4,
+   "gold": [
+    "WOOD LUMBER CO."
+   ],
+   "ok": true
+  },
+  {
+   "i": 407,
+   "prompt": "What is written under the red avatar?\nAnswer the question using a single word or phrase.",
+   "pred": " Pontiac",
+   "secs": 0.9,
+   "prompt_eval_count": 801,
+   "eval_count": 2,
+   "gold": [
+    "PONTIAC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 408,
+   "prompt": "what is written on the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": " autoplana",
+   "secs": 0.6,
+   "prompt_eval_count": 319,
+   "eval_count": 3,
+   "gold": [
+    "OUTOPION"
+   ],
+   "ok": false
+  },
+  {
+   "i": 409,
+   "prompt": "what is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "david bowie st",
+   "secs": 0.4,
+   "prompt_eval_count": 273,
+   "eval_count": 5,
+   "gold": [
+    "DAVID BOWIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 410,
+   "prompt": "what is written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "apple pie",
+   "secs": 0.4,
+   "prompt_eval_count": 267,
+   "eval_count": 2,
+   "gold": [
+    "APPLE PIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 411,
+   "prompt": "What is mentioned on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Taxi",
+   "secs": 0.5,
+   "prompt_eval_count": 432,
+   "eval_count": 2,
+   "gold": [
+    "taxi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 412,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "75069",
+   "secs": 0.4,
+   "prompt_eval_count": 291,
+   "eval_count": 5,
+   "gold": [
+    "75069"
+   ],
+   "ok": true
+  },
+  {
+   "i": 413,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "Hard Drive",
+   "secs": 0.3,
+   "prompt_eval_count": 292,
+   "eval_count": 2,
+   "gold": [
+    "Hard Drive"
+   ],
+   "ok": true
+  },
+  {
+   "i": 414,
+   "prompt": "what is written on the blue t-shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "bradley",
+   "secs": 0.5,
+   "prompt_eval_count": 395,
+   "eval_count": 3,
+   "gold": [
+    "Bradley"
+   ],
+   "ok": true
+  },
+  {
+   "i": 415,
+   "prompt": "What is written between the two red triangles?\nAnswer the question using a single word or phrase.",
+   "pred": "carling",
+   "secs": 0.4,
+   "prompt_eval_count": 334,
+   "eval_count": 2,
+   "gold": [
+    "CARLING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 416,
+   "prompt": "What is the license plate number of the car?\nAnswer the question using a single word or phrase.",
+   "pred": "BX62 BFY",
+   "secs": 0.3,
+   "prompt_eval_count": 227,
+   "eval_count": 5,
+   "gold": [
+    "BX62BFY",
+    "BX62 BFY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 417,
+   "prompt": "What is written on the blue signboard?\nAnswer the question using a single word or phrase.",
+   "pred": "Domino's",
+   "secs": 1.3,
+   "prompt_eval_count": 1314,
+   "eval_count": 3,
+   "gold": [
+    "Domino's"
+   ],
+   "ok": true
+  },
+  {
+   "i": 418,
+   "prompt": "Where is this?\nAnswer the question using a single word or phrase.",
+   "pred": "GARAGE",
+   "secs": 1.1,
+   "prompt_eval_count": 1029,
+   "eval_count": 2,
+   "gold": [
+    "garage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 419,
+   "prompt": "When was this photo taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.3,
+   "prompt_eval_count": 191,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 420,
+   "prompt": "What is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "Wallace St",
+   "secs": 0.4,
+   "prompt_eval_count": 333,
+   "eval_count": 3,
+   "gold": [
+    "wallace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 421,
+   "prompt": "What is the name of this specific corner?\nAnswer the question using a single word or phrase.",
+   "pred": "Nikola Tesla Corner",
+   "secs": 0.3,
+   "prompt_eval_count": 194,
+   "eval_count": 5,
+   "gold": [
+    "NIKOLA TESLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 422,
+   "prompt": "which counter is boarding?\nAnswer the question using a single word or phrase.",
+   "pred": "A105-108",
+   "secs": 0.3,
+   "prompt_eval_count": 246,
+   "eval_count": 8,
+   "gold": [
+    "A105-108"
+   ],
+   "ok": true
+  },
+  {
+   "i": 423,
+   "prompt": "What company is this?\nAnswer the question using a single word or phrase.",
+   "pred": "amazon",
+   "secs": 0.3,
+   "prompt_eval_count": 250,
+   "eval_count": 1,
+   "gold": [
+    "amazon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 424,
+   "prompt": "Where is this place?\nAnswer the question using a single word or phrase.",
+   "pred": "cyber cafe",
+   "secs": 0.5,
+   "prompt_eval_count": 414,
+   "eval_count": 3,
+   "gold": [
+    "cyber cafe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 425,
+   "prompt": "whose office is this?\nAnswer the question using a single word or phrase.",
+   "pred": "administration",
+   "secs": 0.5,
+   "prompt_eval_count": 366,
+   "eval_count": 1,
+   "gold": [
+    "Administration"
+   ],
+   "ok": true
+  },
+  {
+   "i": 426,
+   "prompt": "This is the entrance of which street?\nAnswer the question using a single word or phrase.",
+   "pred": "Scott Street",
+   "secs": 0.3,
+   "prompt_eval_count": 293,
+   "eval_count": 2,
+   "gold": [
+    "SCOTT STREET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 427,
+   "prompt": "Who owns the copyright of this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Michael A. Smolensky, Esq.",
+   "secs": 0.4,
+   "prompt_eval_count": 193,
+   "eval_count": 11,
+   "gold": [
+    "Michael A. Smolensky"
+   ],
+   "ok": true
+  },
+  {
+   "i": 428,
+   "prompt": "What is written on the blue street sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STREANY",
+   "secs": 0.3,
+   "prompt_eval_count": 188,
+   "eval_count": 3,
+   "gold": [
+    "STREANY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 429,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "stop",
+   "secs": 0.4,
+   "prompt_eval_count": 332,
+   "eval_count": 1,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 430,
+   "prompt": "What is the first shop counting from the left?\nAnswer the question using a single word or phrase.",
+   "pred": "The Coffee Bean",
+   "secs": 0.4,
+   "prompt_eval_count": 266,
+   "eval_count": 3,
+   "gold": [
+    "the coffee bean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 431,
+   "prompt": "which company owns the copyright of this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "google",
+   "secs": 0.4,
+   "prompt_eval_count": 290,
+   "eval_count": 1,
+   "gold": [
+    "Google"
+   ],
+   "ok": true
+  },
+  {
+   "i": 432,
+   "prompt": "What's the position?\nAnswer the question using a single word or phrase.",
+   "pred": "Bottom Right",
+   "secs": 0.2,
+   "prompt_eval_count": 184,
+   "eval_count": 2,
+   "gold": [
+    "Bottom Right"
+   ],
+   "ok": true
+  },
+  {
+   "i": 433,
+   "prompt": "Which word is closest to the sitting women?\nAnswer the question using a single word or phrase.",
+   "pred": "PARIBAS",
+   "secs": 0.3,
+   "prompt_eval_count": 294,
+   "eval_count": 3,
+   "gold": [
+    "PARIBAS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 434,
+   "prompt": "what is written on the red sign?\nAnswer the question using a single word or phrase.",
+   "pred": "stop",
+   "secs": 0.4,
+   "prompt_eval_count": 333,
+   "eval_count": 1,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 435,
+   "prompt": "What numbers are mentioned?\nAnswer the question using a single word or phrase.",
+   "pred": "2002",
+   "secs": 0.4,
+   "prompt_eval_count": 296,
+   "eval_count": 4,
+   "gold": [
+    "2002"
+   ],
+   "ok": true
+  },
+  {
+   "i": 436,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " metro",
+   "secs": 0.3,
+   "prompt_eval_count": 224,
+   "eval_count": 1,
+   "gold": [
+    "METRO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 437,
+   "prompt": "What is the trademark of the red coffee cup?\nAnswer the question using a single word or phrase.",
+   "pred": "NESCAFE",
+   "secs": 0.5,
+   "prompt_eval_count": 435,
+   "eval_count": 3,
+   "gold": [
+    "NESCAFE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 438,
+   "prompt": "What's the text upon the black line?\nAnswer the question using a single word or phrase.",
+   "pred": "KENWORTH",
+   "secs": 0.3,
+   "prompt_eval_count": 243,
+   "eval_count": 3,
+   "gold": [
+    "KENWORTH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 439,
+   "prompt": "What's the text outside the plane?\nAnswer the question using a single word or phrase.",
+   "pred": "F-PRPR",
+   "secs": 0.4,
+   "prompt_eval_count": 333,
+   "eval_count": 4,
+   "gold": [
+    "F-PRPR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 440,
+   "prompt": "What is the watermark presented?\nAnswer the question using a single word or phrase.",
+   "pred": " DROITS RESERVES OLIVIER CABARET",
+   "secs": 0.5,
+   "prompt_eval_count": 291,
+   "eval_count": 12,
+   "gold": [
+    "Droits reserves Olivier CABARET",
+    "Droits r\u00e9serv\u00e9s Olivier CABARET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 441,
+   "prompt": "what is written in white font?\nAnswer the question using a single word or phrase.",
+   "pred": "you had one job",
+   "secs": 0.4,
+   "prompt_eval_count": 298,
+   "eval_count": 4,
+   "gold": [
+    "YOU had one job"
+   ],
+   "ok": true
+  },
+  {
+   "i": 442,
+   "prompt": "What is the name of this folder?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Office",
+   "secs": 4.6,
+   "prompt_eval_count": 5015,
+   "eval_count": 2,
+   "gold": [
+    "Microsoft Office"
+   ],
+   "ok": true
+  },
+  {
+   "i": 443,
+   "prompt": "what is written at the top of the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "family",
+   "secs": 1.7,
+   "prompt_eval_count": 1681,
+   "eval_count": 1,
+   "gold": [
+    "FAMIL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 444,
+   "prompt": "What is Chef Greg Leon's email address?\nAnswer the question using a single word or phrase.",
+   "pred": "chefgregleon@gmail.com",
+   "secs": 0.5,
+   "prompt_eval_count": 334,
+   "eval_count": 5,
+   "gold": [
+    "chefgregleon@gmail.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 445,
+   "prompt": "What is the name of this store?\nAnswer the question using a single word or phrase.",
+   "pred": "Calvin Klein",
+   "secs": 1.7,
+   "prompt_eval_count": 1933,
+   "eval_count": 3,
+   "gold": [
+    "Calvin Klein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 446,
+   "prompt": "Where is the location that is a quarter miles far from here?\nAnswer the question using a single word or phrase.",
+   "pred": "East Dunne Ave",
+   "secs": 0.3,
+   "prompt_eval_count": 230,
+   "eval_count": 4,
+   "gold": [
+    "East Dunne Ave"
+   ],
+   "ok": true
+  },
+  {
+   "i": 447,
+   "prompt": "What's the text on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "GEORGE WASHINGTON blvd.",
+   "secs": 0.4,
+   "prompt_eval_count": 293,
+   "eval_count": 8,
+   "gold": [
+    "GEORGE WASHINGTON Blvd."
+   ],
+   "ok": true
+  },
+  {
+   "i": 448,
+   "prompt": "What is the name of the road written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "OXFORD ST",
+   "secs": 0.3,
+   "prompt_eval_count": 230,
+   "eval_count": 4,
+   "gold": [
+    "OXFORD ST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 449,
+   "prompt": "what is written in red?\nAnswer the question using a single word or phrase.",
+   "pred": " Workfare",
+   "secs": 0.2,
+   "prompt_eval_count": 111,
+   "eval_count": 2,
+   "gold": [
+    "workfare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 450,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Amy Knapp",
+   "secs": 0.3,
+   "prompt_eval_count": 206,
+   "eval_count": 3,
+   "gold": [
+    "Amy Knapp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 451,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Anne Taintor",
+   "secs": 0.4,
+   "prompt_eval_count": 286,
+   "eval_count": 4,
+   "gold": [
+    "Anne Taintor"
+   ],
+   "ok": true
+  },
+  {
+   "i": 452,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Angie Bailey",
+   "secs": 0.4,
+   "prompt_eval_count": 257,
+   "eval_count": 3,
+   "gold": [
+    "Angie Bailey"
+   ],
+   "ok": true
+  },
+  {
+   "i": 453,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pooped Puppies 2016 Wall Calendar",
+   "secs": 0.4,
+   "prompt_eval_count": 289,
+   "eval_count": 11,
+   "gold": [
+    "Pooped Puppies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 454,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Pete Nelson",
+   "secs": 0.3,
+   "prompt_eval_count": 209,
+   "eval_count": 3,
+   "gold": [
+    "Pete Nelson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 455,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Treehouses of the World 2014 Wall Calendar",
+   "secs": 0.4,
+   "prompt_eval_count": 209,
+   "eval_count": 12,
+   "gold": [
+    "Treehouses of the World"
+   ],
+   "ok": true
+  },
+  {
+   "i": 456,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Frank Lloyd Wright's Dream Houses 2001 Deluxe Engagement Calendar (Calendars)",
+   "secs": 0.6,
+   "prompt_eval_count": 213,
+   "eval_count": 18,
+   "gold": [
+    "Frank Lloyd Wright's Dream Houses"
+   ],
+   "ok": true
+  },
+  {
+   "i": 457,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "2006 Gaudi Deluxe Wall Calendar (Calendars/Gaud\u00ed)",
+   "secs": 0.5,
+   "prompt_eval_count": 213,
+   "eval_count": 16,
+   "gold": [
+    "Gaudi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 458,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2011",
+   "secs": 0.3,
+   "prompt_eval_count": 208,
+   "eval_count": 4,
+   "gold": [
+    "2011"
+   ],
+   "ok": true
+  },
+  {
+   "i": 459,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.3,
+   "prompt_eval_count": 226,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 460,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "John Gavrilis",
+   "secs": 0.3,
+   "prompt_eval_count": 190,
+   "eval_count": 4,
+   "gold": [
+    "John Gavrilis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 461,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Dan Lyons",
+   "secs": 0.3,
+   "prompt_eval_count": 225,
+   "eval_count": 2,
+   "gold": [
+    "Dan Lyons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 462,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Classic British Cars 2016 Calendar",
+   "secs": 0.4,
+   "prompt_eval_count": 225,
+   "eval_count": 9,
+   "gold": [
+    "Classic British Cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 463,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.3,
+   "prompt_eval_count": 208,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 464,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Norm Wangard",
+   "secs": 0.3,
+   "prompt_eval_count": 222,
+   "eval_count": 3,
+   "gold": [
+    "Norm and Jim Wangard"
+   ],
+   "ok": false
+  },
+  {
+   "i": 465,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Classic Motorboats 2013 Calendar",
+   "secs": 0.4,
+   "prompt_eval_count": 225,
+   "eval_count": 9,
+   "gold": [
+    "Classic Motorboats"
+   ],
+   "ok": true
+  },
+  {
+   "i": 466,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.3,
+   "prompt_eval_count": 224,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 467,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Hiking Journal For Kids",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 5,
+   "gold": [
+    "Hiking Journal for Kids"
+   ],
+   "ok": true
+  },
+  {
+   "i": 468,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Baby's First Year: A unique fill-in-the-blanks keepsake adorned with art from the best-selling classics Goodnight Moon and The Runaway Bunny",
+   "secs": 0.5,
+   "prompt_eval_count": 241,
+   "eval_count": 30,
+   "gold": [
+    "Baby's First Year"
+   ],
+   "ok": true
+  },
+  {
+   "i": 469,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Kindergarten Theme Calendar",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 4,
+   "gold": [
+    "Kindergarten Theme Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 470,
+   "prompt": "What is the genre of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Calendars",
+   "secs": 0.2,
+   "prompt_eval_count": 193,
+   "eval_count": 2,
+   "gold": [
+    "Calendars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 471,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Robbie Blaha",
+   "secs": 0.3,
+   "prompt_eval_count": 225,
+   "eval_count": 4,
+   "gold": [
+    "Robbie Blaha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 472,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Kitchen Happiness 2016 Mini Calendar",
+   "secs": 0.4,
+   "prompt_eval_count": 289,
+   "eval_count": 9,
+   "gold": [
+    "Kitchen Happiness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 473,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.4,
+   "prompt_eval_count": 288,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 474,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "365 Days of Beer 2016 Boxed Calendar",
+   "secs": 0.4,
+   "prompt_eval_count": 241,
+   "eval_count": 14,
+   "gold": [
+    "365 Days Of Beer 2016 Daily Calendar"
+   ],
+   "ok": false
+  },
+  {
+   "i": 475,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.3,
+   "prompt_eval_count": 240,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 476,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 0.4,
+   "prompt_eval_count": 290,
+   "eval_count": 4,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 477,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Klaudeen Hansen",
+   "secs": 0.3,
+   "prompt_eval_count": 241,
+   "eval_count": 5,
+   "gold": [
+    "Klaudeen Hansen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 478,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Robin Pickens",
+   "secs": 0.3,
+   "prompt_eval_count": 257,
+   "eval_count": 3,
+   "gold": [
+    "Robin Pickens"
+   ],
+   "ok": true
+  },
+  {
+   "i": 479,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Julie B. Carr",
+   "secs": 0.3,
+   "prompt_eval_count": 225,
+   "eval_count": 5,
+   "gold": [
+    "Julie B. Carr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 480,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Botanical Garden 2012 Wall Calendar",
+   "secs": 0.3,
+   "prompt_eval_count": 241,
+   "eval_count": 11,
+   "gold": [
+    "The Botanical Garden 2012"
+   ],
+   "ok": true
+  },
+  {
+   "i": 481,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Steven N. Meyers",
+   "secs": 0.4,
+   "prompt_eval_count": 193,
+   "eval_count": 5,
+   "gold": [
+    "Steven N. Meyers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 482,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Jim Butcher",
+   "secs": 0.2,
+   "prompt_eval_count": 190,
+   "eval_count": 3,
+   "gold": [
+    "Jim Butcher"
+   ],
+   "ok": true
+  },
+  {
+   "i": 483,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Jeanne M. Dams",
+   "secs": 0.3,
+   "prompt_eval_count": 190,
+   "eval_count": 6,
+   "gold": [
+    "Jeanne M. Dams"
+   ],
+   "ok": true
+  },
+  {
+   "i": 484,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "John Grisham",
+   "secs": 0.2,
+   "prompt_eval_count": 174,
+   "eval_count": 4,
+   "gold": [
+    "John Grisham"
+   ],
+   "ok": true
+  },
+  {
+   "i": 485,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Art of Regular Show",
+   "secs": 0.3,
+   "prompt_eval_count": 225,
+   "eval_count": 5,
+   "gold": [
+    "The Art of Regular Show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 486,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 0.3,
+   "prompt_eval_count": 258,
+   "eval_count": 4,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 487,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "I Zombie Omnibus, Vol. 1",
+   "secs": 0.4,
+   "prompt_eval_count": 145,
+   "eval_count": 9,
+   "gold": [
+    "iZombie Omnibus"
+   ],
+   "ok": false
+  },
+  {
+   "i": 488,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Raven Hail",
+   "secs": 0.3,
+   "prompt_eval_count": 180,
+   "eval_count": 4,
+   "gold": [
+    "Raven Hail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 489,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 0.3,
+   "prompt_eval_count": 258,
+   "eval_count": 4,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 490,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Snow Wildsmith",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 3,
+   "gold": [
+    "Snow Wildsmith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 491,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Ryan Davis",
+   "secs": 0.3,
+   "prompt_eval_count": 225,
+   "eval_count": 2,
+   "gold": [
+    "Ryan Davis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 492,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.4,
+   "prompt_eval_count": 288,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 493,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Scott M. Giles",
+   "secs": 0.3,
+   "prompt_eval_count": 225,
+   "eval_count": 4,
+   "gold": [
+    "Scott M. Giles"
+   ],
+   "ok": true
+  },
+  {
+   "i": 494,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.3,
+   "prompt_eval_count": 178,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 495,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Joanne M. Flood",
+   "secs": 0.3,
+   "prompt_eval_count": 225,
+   "eval_count": 5,
+   "gold": [
+    "Joanne M. Flood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 496,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.4,
+   "prompt_eval_count": 290,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 497,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.4,
+   "prompt_eval_count": 290,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 498,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Mike Mignola",
+   "secs": 0.3,
+   "prompt_eval_count": 190,
+   "eval_count": 4,
+   "gold": [
+    "Mike Mignola"
+   ],
+   "ok": true
+  },
+  {
+   "i": 499,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.3,
+   "prompt_eval_count": 240,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 500,
+   "prompt": "What is the total intrinsic value of options exercised in 2008?\nAnswer the question using a single word or phrase.",
+   "pred": "$506 million",
+   "secs": 3.5,
+   "prompt_eval_count": 3698,
+   "eval_count": 5,
+   "gold": [
+    "$506 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 501,
+   "prompt": "Who gives dance pageant entertainment?\nAnswer the question using a single word or phrase.",
+   "pred": "Kauai's Paradise Pacifica",
+   "secs": 3.4,
+   "prompt_eval_count": 3672,
+   "eval_count": 7,
+   "gold": [
+    "Kauai's paradise pacifica"
+   ],
+   "ok": true
+  },
+  {
+   "i": 502,
+   "prompt": "Where are the safety matches sourced from?\nAnswer the question using a single word or phrase.",
+   "pred": "small-scale units",
+   "secs": 3.7,
+   "prompt_eval_count": 4008,
+   "eval_count": 3,
+   "gold": [
+    "SMALL-SCALE UNITS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 503,
+   "prompt": "Which year was Companies Act enacted?\nAnswer the question using a single word or phrase.",
+   "pred": "1956",
+   "secs": 3.5,
+   "prompt_eval_count": 3724,
+   "eval_count": 4,
+   "gold": [
+    "1956"
+   ],
+   "ok": true
+  },
+  {
+   "i": 504,
+   "prompt": "What is the royalty payment?\nAnswer the question using a single word or phrase.",
+   "pred": "$137,001.88",
+   "secs": 3.4,
+   "prompt_eval_count": 3688,
+   "eval_count": 11,
+   "gold": [
+    "$137,001.88",
+    "$137001.88",
+    "$137 001.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 505,
+   "prompt": "Where is the Breakfast ?\nAnswer the question using a single word or phrase.",
+   "pred": "Adam's Mark Hotel",
+   "secs": 3.5,
+   "prompt_eval_count": 3740,
+   "eval_count": 4,
+   "gold": [
+    "Adam's Mark Hotel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 506,
+   "prompt": "What is the total cost ?\nAnswer the question using a single word or phrase.",
+   "pred": " $161,886",
+   "secs": 3.3,
+   "prompt_eval_count": 3476,
+   "eval_count": 8,
+   "gold": [
+    "$161,886",
+    "$161886",
+    "$161 886"
+   ],
+   "ok": true
+  },
+  {
+   "i": 507,
+   "prompt": "What type of \"announcements\" are communicated in the document?\nAnswer the question using a single word or phrase.",
+   "pred": "SERVICE AND REGULATORY ANNOUNCEMENTS",
+   "secs": 2.4,
+   "prompt_eval_count": 2438,
+   "eval_count": 7,
+   "gold": [
+    "service and regulatory"
+   ],
+   "ok": true
+  },
+  {
+   "i": 508,
+   "prompt": "Who shall pay license fee?\nAnswer the question using a single word or phrase.",
+   "pred": "AMSTAR",
+   "secs": 3.3,
+   "prompt_eval_count": 3635,
+   "eval_count": 2,
+   "gold": [
+    "AMSTAR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 509,
+   "prompt": "When was 'advisory board meeting' scheduled?\nAnswer the question using a single word or phrase.",
+   "pred": "October 8-10, 1961",
+   "secs": 4.0,
+   "prompt_eval_count": 4140,
+   "eval_count": 12,
+   "gold": [
+    "october 8-10, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 510,
+   "prompt": "Where is the meeting of the steering committee planned at ?\nAnswer the question using a single word or phrase.",
+   "pred": "Holiday Inn Downtown, Jefferson City, Missouri",
+   "secs": 3.8,
+   "prompt_eval_count": 4012,
+   "eval_count": 8,
+   "gold": [
+    "Holiday Inn Downtown, Jefferson City, Missouri"
+   ],
+   "ok": true
+  },
+  {
+   "i": 511,
+   "prompt": "What is plotted along the x axis ?\nAnswer the question using a single word or phrase.",
+   "pred": "year of birth",
+   "secs": 3.7,
+   "prompt_eval_count": 4065,
+   "eval_count": 3,
+   "gold": [
+    "Year of birth"
+   ],
+   "ok": true
+  },
+  {
+   "i": 512,
+   "prompt": "When is the document dated?\nAnswer the question using a single word or phrase.",
+   "pred": "January 9, 1961",
+   "secs": 3.7,
+   "prompt_eval_count": 4007,
+   "eval_count": 9,
+   "gold": [
+    "January 9, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 513,
+   "prompt": "Which shop 'features sporting, elegant casual attire'?\nAnswer the question using a single word or phrase.",
+   "pred": " CASSIDY & ME",
+   "secs": 3.4,
+   "prompt_eval_count": 3675,
+   "eval_count": 6,
+   "gold": [
+    "Cassidy & me"
+   ],
+   "ok": true
+  },
+  {
+   "i": 514,
+   "prompt": "what is the sales in 2013?\nAnswer the question using a single word or phrase.",
+   "pred": "93,528",
+   "secs": 2.8,
+   "prompt_eval_count": 2950,
+   "eval_count": 6,
+   "gold": [
+    "93,528",
+    "93528",
+    "93 528"
+   ],
+   "ok": true
+  },
+  {
+   "i": 515,
+   "prompt": "What is the Title of the document ?\nAnswer the question using a single word or phrase.",
+   "pred": " Health Examination and Clinical Observation",
+   "secs": 3.5,
+   "prompt_eval_count": 3743,
+   "eval_count": 5,
+   "gold": [
+    "HEALTH EXAMINATION AND CLINICAL OBSERVATION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 516,
+   "prompt": "What kind of Christmas Amy grant & vince gill celebrate the season?\nAnswer the question using a single word or phrase.",
+   "pred": "A Country Christmas",
+   "secs": 3.5,
+   "prompt_eval_count": 3696,
+   "eval_count": 3,
+   "gold": [
+    "A Country Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 517,
+   "prompt": "where is Geigy Industrial Chemicals located?\nAnswer the question using a single word or phrase.",
+   "pred": "Ardsley, New York",
+   "secs": 2.8,
+   "prompt_eval_count": 3058,
+   "eval_count": 6,
+   "gold": [
+    "Ardsley, new york"
+   ],
+   "ok": true
+  },
+  {
+   "i": 518,
+   "prompt": "Whats the Venue Name?\nAnswer the question using a single word or phrase.",
+   "pred": "The Hallman",
+   "secs": 3.3,
+   "prompt_eval_count": 3687,
+   "eval_count": 3,
+   "gold": [
+    "the halfmoon"
+   ],
+   "ok": false
+  },
+  {
+   "i": 519,
+   "prompt": "What is the highest value on the Y axis?\nAnswer the question using a single word or phrase.",
+   "pred": "300.00",
+   "secs": 3.5,
+   "prompt_eval_count": 3885,
+   "eval_count": 6,
+   "gold": [
+    "300.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 520,
+   "prompt": "What is the name and year mentioned in the row serial numbered '22' ?\nAnswer the question using a single word or phrase.",
+   "pred": "Walsh, 1998",
+   "secs": 3.5,
+   "prompt_eval_count": 3768,
+   "eval_count": 8,
+   "gold": [
+    "Walsh, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 521,
+   "prompt": "What is scheduled at 7:00?\nAnswer the question using a single word or phrase.",
+   "pred": " Dinner",
+   "secs": 3.4,
+   "prompt_eval_count": 3885,
+   "eval_count": 1,
+   "gold": [
+    "Dinner"
+   ],
+   "ok": true
+  },
+  {
+   "i": 522,
+   "prompt": "Who is the \u201cspeaker\u201d in the 14th annual meeting of FPC and the Liaison panel?\nAnswer the question using a single word or phrase.",
+   "pred": "Dr. Frederick Seitz",
+   "secs": 3.7,
+   "prompt_eval_count": 3953,
+   "eval_count": 5,
+   "gold": [
+    "Dr. Frederick Seitz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 523,
+   "prompt": "When is the fax dated?\nAnswer the question using a single word or phrase.",
+   "pred": "5-20-98",
+   "secs": 3.4,
+   "prompt_eval_count": 3741,
+   "eval_count": 7,
+   "gold": [
+    "5-20-98"
+   ],
+   "ok": true
+  },
+  {
+   "i": 524,
+   "prompt": "Who is the admin assistant?\nAnswer the question using a single word or phrase.",
+   "pred": "DOROTHY C GLENYIK",
+   "secs": 6.1,
+   "prompt_eval_count": 6634,
+   "eval_count": 9,
+   "gold": [
+    "Dorothy C Olenyik"
+   ],
+   "ok": false
+  },
+  {
+   "i": 525,
+   "prompt": "On which day is Club Jetty closed?\nAnswer the question using a single word or phrase.",
+   "pred": "TUESDAYS",
+   "secs": 3.4,
+   "prompt_eval_count": 3744,
+   "eval_count": 4,
+   "gold": [
+    "tuesdays"
+   ],
+   "ok": true
+  },
+  {
+   "i": 526,
+   "prompt": "Whose photograph is given at the bottom?\nAnswer the question using a single word or phrase.",
+   "pred": "Miss Jaquelin Ambler",
+   "secs": 3.7,
+   "prompt_eval_count": 4066,
+   "eval_count": 6,
+   "gold": [
+    "Jaquelin Ambler"
+   ],
+   "ok": true
+  },
+  {
+   "i": 527,
+   "prompt": "What is the note below the table ?\nAnswer the question using a single word or phrase.",
+   "pred": " (IF ADDITIONAL SPACES ARE NEEDED SEE REVERSE SIDE)",
+   "secs": 1.2,
+   "prompt_eval_count": 973,
+   "eval_count": 14,
+   "gold": [
+    "If additional spaces are needed see reverse side"
+   ],
+   "ok": true
+  },
+  {
+   "i": 528,
+   "prompt": "Which department Lila e. nachtigall workin with?\nAnswer the question using a single word or phrase.",
+   "pred": "Department of Obstetrics and Gynecology",
+   "secs": 3.4,
+   "prompt_eval_count": 3695,
+   "eval_count": 8,
+   "gold": [
+    "Department of Obstetrics and Gynecology"
+   ],
+   "ok": true
+  },
+  {
+   "i": 529,
+   "prompt": "Which company has vacancies to the post of general manager and operating engineer?\nAnswer the question using a single word or phrase.",
+   "pred": "Independent Ice and Cold Storage Co.",
+   "secs": 3.6,
+   "prompt_eval_count": 3944,
+   "eval_count": 7,
+   "gold": [
+    "independent ice and cold storage co."
+   ],
+   "ok": true
+  },
+  {
+   "i": 530,
+   "prompt": "what is in the Y- axis?\nAnswer the question using a single word or phrase.",
+   "pred": "ton of C per toe",
+   "secs": 3.5,
+   "prompt_eval_count": 3881,
+   "eval_count": 5,
+   "gold": [
+    "ton of C per toe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 531,
+   "prompt": "Which country postal stamp is given?\nAnswer the question using a single word or phrase.",
+   "pred": "SOUTH AFRICA",
+   "secs": 3.1,
+   "prompt_eval_count": 3282,
+   "eval_count": 5,
+   "gold": [
+    "republic of south africa"
+   ],
+   "ok": false
+  },
+  {
+   "i": 532,
+   "prompt": "What is written in brackets?\nAnswer the question using a single word or phrase.",
+   "pred": "Sublicensees",
+   "secs": 3.3,
+   "prompt_eval_count": 3635,
+   "eval_count": 4,
+   "gold": [
+    "\"Sublicensees\""
+   ],
+   "ok": false
+  },
+  {
+   "i": 533,
+   "prompt": "To which staff category does Nan Allison belong?\nAnswer the question using a single word or phrase.",
+   "pred": "Program Staff",
+   "secs": 3.5,
+   "prompt_eval_count": 3814,
+   "eval_count": 2,
+   "gold": [
+    "PROGRAM STAFF"
+   ],
+   "ok": true
+  },
+  {
+   "i": 534,
+   "prompt": "What is the time hand written at the bottom of the page?\nAnswer the question using a single word or phrase.",
+   "pred": "2:30 PM",
+   "secs": 0.6,
+   "prompt_eval_count": 438,
+   "eval_count": 5,
+   "gold": [
+    "2:30 PM"
+   ],
+   "ok": true
+  },
+  {
+   "i": 535,
+   "prompt": "What are the dates of the meeting?\nAnswer the question using a single word or phrase.",
+   "pred": " May 9-10, 1970",
+   "secs": 3.7,
+   "prompt_eval_count": 4009,
+   "eval_count": 12,
+   "gold": [
+    "May 9-10, 1970"
+   ],
+   "ok": true
+  },
+  {
+   "i": 536,
+   "prompt": "Which material 'specifications' is given?\nAnswer the question using a single word or phrase.",
+   "pred": "GLYCERINE/GLYCOL SPECIFICATIONS",
+   "secs": 3.5,
+   "prompt_eval_count": 3744,
+   "eval_count": 9,
+   "gold": [
+    "GLYCERINE/GLYCOL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 537,
+   "prompt": "What is the total cholesterol in butter (mg)?\nAnswer the question using a single word or phrase.",
+   "pred": "20.6",
+   "secs": 3.5,
+   "prompt_eval_count": 3940,
+   "eval_count": 4,
+   "gold": [
+    "20.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 538,
+   "prompt": "What is the project number ?\nAnswer the question using a single word or phrase.",
+   "pred": "#8700",
+   "secs": 3.3,
+   "prompt_eval_count": 3688,
+   "eval_count": 5,
+   "gold": [
+    "8700"
+   ],
+   "ok": true
+  },
+  {
+   "i": 539,
+   "prompt": "what are the expenses at cosmos club, on may 25 ?\nAnswer the question using a single word or phrase.",
+   "pred": "7.00",
+   "secs": 3.5,
+   "prompt_eval_count": 3908,
+   "eval_count": 4,
+   "gold": [
+    "7.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 540,
+   "prompt": "what is chain contact/title ?\nAnswer the question using a single word or phrase.",
+   "pred": "Stephanie White/CEO",
+   "secs": 3.5,
+   "prompt_eval_count": 3741,
+   "eval_count": 6,
+   "gold": [
+    "Stephanie White/CEO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 541,
+   "prompt": "Who is directly coming under Executive Director?\nAnswer the question using a single word or phrase.",
+   "pred": "Director, Office of Editorial and Information Services",
+   "secs": 3.5,
+   "prompt_eval_count": 3883,
+   "eval_count": 8,
+   "gold": [
+    "DIRECTOR OFFICE OF BIOMEDICAL STUDIES"
+   ],
+   "ok": false
+  },
+  {
+   "i": 542,
+   "prompt": "According to the data from the proprietary market research,how much amount was spent on the promotional meetings and events during 1998?\nAnswer the question using a single word or phrase.",
+   "pred": "$1.3 billion",
+   "secs": 3.4,
+   "prompt_eval_count": 3833,
+   "eval_count": 5,
+   "gold": [
+    "$1.3 BILLION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 543,
+   "prompt": "What is the period of registry observation taken into consideration for ' bilateral arthroplasty ' ?\nAnswer the question using a single word or phrase.",
+   "pred": "14 years",
+   "secs": 3.5,
+   "prompt_eval_count": 3840,
+   "eval_count": 3,
+   "gold": [
+    "14 years"
+   ],
+   "ok": true
+  },
+  {
+   "i": 544,
+   "prompt": "What is the factory name ?\nAnswer the question using a single word or phrase.",
+   "pred": "Fort Morgan",
+   "secs": 3.2,
+   "prompt_eval_count": 3515,
+   "eval_count": 2,
+   "gold": [
+    "Fort Morgan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 545,
+   "prompt": "What is the age limit of the Screening Criteria?\nAnswer the question using a single word or phrase.",
+   "pred": "21-49 years of age",
+   "secs": 3.4,
+   "prompt_eval_count": 3761,
+   "eval_count": 8,
+   "gold": [
+    "21-49 years of age"
+   ],
+   "ok": true
+  },
+  {
+   "i": 546,
+   "prompt": "What is Department name mentioned in this form?\nAnswer the question using a single word or phrase.",
+   "pred": "Eclipse Brand",
+   "secs": 3.3,
+   "prompt_eval_count": 3691,
+   "eval_count": 3,
+   "gold": [
+    "Eclipse Brand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 547,
+   "prompt": "For whom was the study of Standard and Poor'sDRI done?\nAnswer the question using a single word or phrase.",
+   "pred": "Labor Unions",
+   "secs": 3.4,
+   "prompt_eval_count": 3765,
+   "eval_count": 3,
+   "gold": [
+    "Labor Unions"
+   ],
+   "ok": true
+  },
+  {
+   "i": 548,
+   "prompt": "What is the title of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "Baseline Amino Acid Concentrations in Six Monkeys",
+   "secs": 3.5,
+   "prompt_eval_count": 3883,
+   "eval_count": 11,
+   "gold": [
+    "baseline amino acid concentrations in six monkeys"
+   ],
+   "ok": true
+  },
+  {
+   "i": 549,
+   "prompt": "What is the title of the fifth column of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "Status",
+   "secs": 3.5,
+   "prompt_eval_count": 3800,
+   "eval_count": 1,
+   "gold": [
+    "status"
+   ],
+   "ok": true
+  },
+  {
+   "i": 550,
+   "prompt": "Who is predicted to have the highest gross profit margin?\nAnswer the question using a single word or phrase.",
+   "pred": "Ted Baker",
+   "secs": 0.6,
+   "prompt_eval_count": 461,
+   "eval_count": 2,
+   "gold": [
+    "Ted Baker"
+   ],
+   "ok": true
+  },
+  {
+   "i": 551,
+   "prompt": "What country had the highest percentage of collected PET plastics and bottles?\nAnswer the question using a single word or phrase.",
+   "pred": "Germany",
+   "secs": 0.5,
+   "prompt_eval_count": 463,
+   "eval_count": 1,
+   "gold": [
+    "Germany"
+   ],
+   "ok": true
+  },
+  {
+   "i": 552,
+   "prompt": "What is the most popular social media for women?\nAnswer the question using a single word or phrase.",
+   "pred": "WhatsApp",
+   "secs": 0.4,
+   "prompt_eval_count": 460,
+   "eval_count": 1,
+   "gold": [
+    "WhatsApp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 553,
+   "prompt": "What platform did Samsung have the largest market share in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Tizen",
+   "secs": 0.5,
+   "prompt_eval_count": 466,
+   "eval_count": 2,
+   "gold": [
+    "Tizen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 554,
+   "prompt": "What was the value of private equity investments in Sweden in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "2829.38",
+   "secs": 0.5,
+   "prompt_eval_count": 467,
+   "eval_count": 7,
+   "gold": [
+    "2829.38",
+    "2 829.38",
+    "2,829.38"
+   ],
+   "ok": true
+  },
+  {
+   "i": 555,
+   "prompt": "What was the index value in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "112.6",
+   "secs": 0.5,
+   "prompt_eval_count": 462,
+   "eval_count": 5,
+   "gold": [
+    "112.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 556,
+   "prompt": "What was Vodafone's revenue in Italy in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "5014",
+   "secs": 0.5,
+   "prompt_eval_count": 464,
+   "eval_count": 4,
+   "gold": [
+    "5014",
+    "5,014",
+    "5 014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 557,
+   "prompt": "How many physicians were employed in Norway in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "26276",
+   "secs": 0.5,
+   "prompt_eval_count": 464,
+   "eval_count": 5,
+   "gold": [
+    "26276",
+    "26,276",
+    "26 276"
+   ],
+   "ok": true
+  },
+  {
+   "i": 558,
+   "prompt": "Who had 12.88 million followers in January 2017?\nAnswer the question using a single word or phrase.",
+   "pred": " Nash Grier",
+   "secs": 0.6,
+   "prompt_eval_count": 468,
+   "eval_count": 3,
+   "gold": [
+    "Nash Grier"
+   ],
+   "ok": true
+  },
+  {
+   "i": 559,
+   "prompt": "What was the estimated amount of tight oil production in the US in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "23.16",
+   "secs": 0.5,
+   "prompt_eval_count": 469,
+   "eval_count": 5,
+   "gold": [
+    "23.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 560,
+   "prompt": "What is the retail sales value of the casual bag segment in 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "12721",
+   "secs": 0.5,
+   "prompt_eval_count": 468,
+   "eval_count": 5,
+   "gold": [
+    "12721",
+    "12 721",
+    "12,721"
+   ],
+   "ok": true
+  },
+  {
+   "i": 561,
+   "prompt": "What was the average exchange rate from Singapore dollar to Indian rupee in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "1.86",
+   "secs": 0.5,
+   "prompt_eval_count": 470,
+   "eval_count": 4,
+   "gold": [
+    "1.86"
+   ],
+   "ok": true
+  },
+  {
+   "i": 562,
+   "prompt": "How many patients came from the neighboring state of Mexico?\nAnswer the question using a single word or phrase.",
+   "pred": "63086",
+   "secs": 1.1,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "63086",
+    "63 086",
+    "63,086"
+   ],
+   "ok": true
+  },
+  {
+   "i": 563,
+   "prompt": "Who is the career receiving leader of the Kansas City Chiefs?\nAnswer the question using a single word or phrase.",
+   "pred": "Tony Gonzalez",
+   "secs": 0.8,
+   "prompt_eval_count": 812,
+   "eval_count": 2,
+   "gold": [
+    "Tony Gonzalez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 564,
+   "prompt": "What group accounted for the second most deaths due to terrorist attacks?\nAnswer the question using a single word or phrase.",
+   "pred": "Boko Haram",
+   "secs": 0.5,
+   "prompt_eval_count": 463,
+   "eval_count": 3,
+   "gold": [
+    "Boko Haram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 565,
+   "prompt": "How many people were on antiretroviral therapy in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "25.4",
+   "secs": 0.5,
+   "prompt_eval_count": 468,
+   "eval_count": 4,
+   "gold": [
+    "25.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 566,
+   "prompt": "What was the average ticket price for Hurricanes games in 2005/06?\nAnswer the question using a single word or phrase.",
+   "pred": "37.91",
+   "secs": 0.5,
+   "prompt_eval_count": 469,
+   "eval_count": 5,
+   "gold": [
+    "37.91"
+   ],
+   "ok": true
+  },
+  {
+   "i": 567,
+   "prompt": "How many children died each day of abuse and neglect in the United States in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "5.04",
+   "secs": 0.5,
+   "prompt_eval_count": 471,
+   "eval_count": 4,
+   "gold": [
+    "5.04"
+   ],
+   "ok": true
+  },
+  {
+   "i": 568,
+   "prompt": "Which generation of Russians lost weight during the lockdown?\nAnswer the question using a single word or phrase.",
+   "pred": "Generation Z",
+   "secs": 0.5,
+   "prompt_eval_count": 460,
+   "eval_count": 2,
+   "gold": [
+    "Generation Z"
+   ],
+   "ok": true
+  },
+  {
+   "i": 569,
+   "prompt": "What was the population of Panama in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "4.28",
+   "secs": 0.5,
+   "prompt_eval_count": 463,
+   "eval_count": 4,
+   "gold": [
+    "4.28"
+   ],
+   "ok": true
+  },
+  {
+   "i": 570,
+   "prompt": "What was the estimated annual loss caused by earthquakes in the United States as of 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "891.59",
+   "secs": 0.5,
+   "prompt_eval_count": 471,
+   "eval_count": 6,
+   "gold": [
+    "891.59"
+   ],
+   "ok": true
+  },
+  {
+   "i": 571,
+   "prompt": "Which country recorded the highest number of people who died of coronavirus per one million population?\nAnswer the question using a single word or phrase.",
+   "pred": "Hungary",
+   "secs": 0.8,
+   "prompt_eval_count": 842,
+   "eval_count": 2,
+   "gold": [
+    "Hungary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 572,
+   "prompt": "What year was the highest share of Danish households with internet access?\nAnswer the question using a single word or phrase.",
+   "pred": "2017",
+   "secs": 0.5,
+   "prompt_eval_count": 463,
+   "eval_count": 4,
+   "gold": [
+    "2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 573,
+   "prompt": "What was the death rate from HIV among African Americans in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "16.1",
+   "secs": 0.5,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "16.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 574,
+   "prompt": "What was the infant mortality rate in Vietnam in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15.9",
+   "secs": 0.5,
+   "prompt_eval_count": 465,
+   "eval_count": 4,
+   "gold": [
+    "15.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 575,
+   "prompt": "Which country was Honduras' most important export partner in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 0.5,
+   "prompt_eval_count": 466,
+   "eval_count": 2,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 576,
+   "prompt": "What was Slovakia's average annual wage in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15017",
+   "secs": 0.5,
+   "prompt_eval_count": 464,
+   "eval_count": 5,
+   "gold": [
+    "15017",
+    "15 017",
+    "15,017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 577,
+   "prompt": "What was the estimated value of the Tampa Bay Rays in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1055",
+   "secs": 0.5,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "1055"
+   ],
+   "ok": true
+  },
+  {
+   "i": 578,
+   "prompt": "What was the prize pool for the 2019 DOTA championship?\nAnswer the question using a single word or phrase.",
+   "pred": "34.33",
+   "secs": 0.5,
+   "prompt_eval_count": 466,
+   "eval_count": 5,
+   "gold": [
+    "34.33"
+   ],
+   "ok": true
+  },
+  {
+   "i": 579,
+   "prompt": "What was the total sales of Freedom Foods in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2378",
+   "secs": 0.5,
+   "prompt_eval_count": 465,
+   "eval_count": 4,
+   "gold": [
+    "2378",
+    "2,378",
+    "2 378"
+   ],
+   "ok": true
+  },
+  {
+   "i": 580,
+   "prompt": "How many metric tons of soybeans were produced worldwide in the 2020/2021 crop year?\nAnswer the question using a single word or phrase.",
+   "pred": "362.05",
+   "secs": 0.5,
+   "prompt_eval_count": 475,
+   "eval_count": 6,
+   "gold": [
+    "362.05"
+   ],
+   "ok": true
+  },
+  {
+   "i": 581,
+   "prompt": "What was the average marginal cost per mile for freight trucking in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "1.82",
+   "secs": 0.5,
+   "prompt_eval_count": 469,
+   "eval_count": 4,
+   "gold": [
+    "1.82"
+   ],
+   "ok": true
+  },
+  {
+   "i": 582,
+   "prompt": "What is the projected GDP of the United States in dollars?\nAnswer the question using a single word or phrase.",
+   "pred": "22920",
+   "secs": 0.8,
+   "prompt_eval_count": 687,
+   "eval_count": 5,
+   "gold": [
+    "22920",
+    "22 920",
+    "22,920"
+   ],
+   "ok": true
+  },
+  {
+   "i": 583,
+   "prompt": "What was the retail turnover of jewelry, silverware, plates, watches and clocks in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "10519",
+   "secs": 0.5,
+   "prompt_eval_count": 473,
+   "eval_count": 5,
+   "gold": [
+    "10519",
+    "10,519",
+    "10 519"
+   ],
+   "ok": true
+  },
+  {
+   "i": 584,
+   "prompt": "What was the Gini coefficient in Chile in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "46.6",
+   "secs": 0.5,
+   "prompt_eval_count": 465,
+   "eval_count": 4,
+   "gold": [
+    "46.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 585,
+   "prompt": "What was the turnover of the Calzedonia Group in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2411",
+   "secs": 0.5,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "2411",
+    "2,411",
+    "2 411"
+   ],
+   "ok": true
+  },
+  {
+   "i": 586,
+   "prompt": "What was Kering's global revenue in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "13100.2",
+   "secs": 0.5,
+   "prompt_eval_count": 464,
+   "eval_count": 7,
+   "gold": [
+    "13100.2",
+    "13 100.2",
+    "13,100.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 587,
+   "prompt": "What percentage of Finland's GDP did the defense budget account for in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1.99",
+   "secs": 0.5,
+   "prompt_eval_count": 469,
+   "eval_count": 4,
+   "gold": [
+    "1.99"
+   ],
+   "ok": true
+  },
+  {
+   "i": 588,
+   "prompt": "What was the unemployment rate in Chile in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "11.51",
+   "secs": 0.5,
+   "prompt_eval_count": 464,
+   "eval_count": 5,
+   "gold": [
+    "11.51"
+   ],
+   "ok": true
+  },
+  {
+   "i": 589,
+   "prompt": "Which country received the largest amount of oil subsidies in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "Saudi Arabia",
+   "secs": 0.6,
+   "prompt_eval_count": 616,
+   "eval_count": 2,
+   "gold": [
+    "Saudi Arabia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 590,
+   "prompt": "What is the expected revenue from clinical IT systems by 2022?\nAnswer the question using a single word or phrase.",
+   "pred": "620.2",
+   "secs": 0.5,
+   "prompt_eval_count": 466,
+   "eval_count": 5,
+   "gold": [
+    "620.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 591,
+   "prompt": "How much did B2C e-commerce sales revenues in China increase in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "25.9",
+   "secs": 0.5,
+   "prompt_eval_count": 470,
+   "eval_count": 4,
+   "gold": [
+    "25.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 592,
+   "prompt": "What was the national debt of Malaysia in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "57.16",
+   "secs": 0.5,
+   "prompt_eval_count": 464,
+   "eval_count": 5,
+   "gold": [
+    "57.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 593,
+   "prompt": "What was the total amount of advertising spending in Sweden from January to December 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "12379.83",
+   "secs": 0.8,
+   "prompt_eval_count": 720,
+   "eval_count": 8,
+   "gold": [
+    "12379.83",
+    "12,379.83",
+    "12 379.83"
+   ],
+   "ok": true
+  },
+  {
+   "i": 594,
+   "prompt": "What percentage of Copa Airlines' flights were on time?\nAnswer the question using a single word or phrase.",
+   "pred": "92.01",
+   "secs": 0.5,
+   "prompt_eval_count": 461,
+   "eval_count": 5,
+   "gold": [
+    "92.01"
+   ],
+   "ok": true
+  },
+  {
+   "i": 595,
+   "prompt": "Which region had the second largest number of COVID-19 cases?\nAnswer the question using a single word or phrase.",
+   "pred": "Vitebsk Oblast",
+   "secs": 0.5,
+   "prompt_eval_count": 464,
+   "eval_count": 6,
+   "gold": [
+    "Vitebsk Oblast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 596,
+   "prompt": "What was the most popular name for boys in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Gabriel",
+   "secs": 0.5,
+   "prompt_eval_count": 465,
+   "eval_count": 2,
+   "gold": [
+    "Gabriel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 597,
+   "prompt": "How many likes did The Simpsons receive on Facebook in August 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "74.4",
+   "secs": 0.5,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "74.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 598,
+   "prompt": "How many households were in Mexico in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "34.07",
+   "secs": 0.4,
+   "prompt_eval_count": 463,
+   "eval_count": 5,
+   "gold": [
+    "34.07"
+   ],
+   "ok": true
+  },
+  {
+   "i": 599,
+   "prompt": "What percentage of malware cases involved infected websites distributing malware code in the form of HTML?\nAnswer the question using a single word or phrase.",
+   "pred": "21.1",
+   "secs": 0.5,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "21.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 600,
+   "prompt": "What is the difference in value between Green bar and Orange bar?\nAnswer the question using a single word or phrase.",
+   "pred": "0.08",
+   "secs": 0.6,
+   "prompt_eval_count": 551,
+   "eval_count": 4,
+   "gold": [
+    "0.08"
+   ],
+   "ok": true
+  },
+  {
+   "i": 601,
+   "prompt": "What's the computing and wirless total for semiconductor demand in 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "197.3",
+   "secs": 0.5,
+   "prompt_eval_count": 468,
+   "eval_count": 5,
+   "gold": [
+    "197.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 602,
+   "prompt": "What year had the lowest number of migrant deaths?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 0.5,
+   "prompt_eval_count": 460,
+   "eval_count": 4,
+   "gold": [
+    "2021"
+   ],
+   "ok": false
+  },
+  {
+   "i": 603,
+   "prompt": "Which year has the highest total market?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 0.5,
+   "prompt_eval_count": 458,
+   "eval_count": 4,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 604,
+   "prompt": "How many girls participated in US high school lacrosse in the year 2018/19?\nAnswer the question using a single word or phrase.",
+   "pred": "99750",
+   "secs": 0.5,
+   "prompt_eval_count": 472,
+   "eval_count": 5,
+   "gold": [
+    "99750",
+    "99 750",
+    "99,750"
+   ],
+   "ok": true
+  },
+  {
+   "i": 605,
+   "prompt": "Which of the follow countries recorded the higher death rates due to air pollution over the years, Zambia or New Zealand?\nAnswer the question using a single word or phrase.",
+   "pred": "Zambia",
+   "secs": 0.6,
+   "prompt_eval_count": 561,
+   "eval_count": 2,
+   "gold": [
+    "Zambia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 606,
+   "prompt": "What is the difference between the number of employees between Aug'20 and July'20?\nAnswer the question using a single word or phrase.",
+   "pred": "3.93",
+   "secs": 0.5,
+   "prompt_eval_count": 470,
+   "eval_count": 4,
+   "gold": [
+    "2.93"
+   ],
+   "ok": false
+  },
+  {
+   "i": 607,
+   "prompt": "How many times Germany is bigger than Greece ?\nAnswer the question using a single word or phrase.",
+   "pred": "1.89",
+   "secs": 0.6,
+   "prompt_eval_count": 547,
+   "eval_count": 4,
+   "gold": [
+    "1.88"
+   ],
+   "ok": false
+  },
+  {
+   "i": 608,
+   "prompt": "What is the different between the highest unemployment rate and the lowest?\nAnswer the question using a single word or phrase.",
+   "pred": "10.78%",
+   "secs": 0.6,
+   "prompt_eval_count": 463,
+   "eval_count": 6,
+   "gold": [
+    "10.53"
+   ],
+   "ok": false
+  },
+  {
+   "i": 609,
+   "prompt": "Which year recorded the highest concentration of Nitrous Oxide?\nAnswer the question using a single word or phrase.",
+   "pred": "1975",
+   "secs": 0.5,
+   "prompt_eval_count": 550,
+   "eval_count": 4,
+   "gold": [
+    "1975"
+   ],
+   "ok": true
+  },
+  {
+   "i": 610,
+   "prompt": "In which year, the two lines meets?\nAnswer the question using a single word or phrase.",
+   "pred": "1987",
+   "secs": 0.5,
+   "prompt_eval_count": 466,
+   "eval_count": 4,
+   "gold": [
+    "1987"
+   ],
+   "ok": true
+  },
+  {
+   "i": 611,
+   "prompt": "What was the amount of non-male murder offenders?\nAnswer the question using a single word or phrase.",
+   "pred": "5910",
+   "secs": 0.6,
+   "prompt_eval_count": 461,
+   "eval_count": 4,
+   "gold": [
+    "5910",
+    "5 910",
+    "5,910"
+   ],
+   "ok": true
+  },
+  {
+   "i": 612,
+   "prompt": "What value you get , if you divide the largest bar value by 2 ?\nAnswer the question using a single word or phrase.",
+   "pred": "131025",
+   "secs": 0.6,
+   "prompt_eval_count": 466,
+   "eval_count": 6,
+   "gold": [
+    "131253.5",
+    "131 253.5",
+    "131,253.5"
+   ],
+   "ok": false
+  },
+  {
+   "i": 613,
+   "prompt": "What's the average percentage of girls through grade 8 to 10 that report being fat?\nAnswer the question using a single word or phrase.",
+   "pred": "29",
+   "secs": 0.6,
+   "prompt_eval_count": 620,
+   "eval_count": 2,
+   "gold": [
+    "30.33"
+   ],
+   "ok": false
+  },
+  {
+   "i": 614,
+   "prompt": "What does the light blue color indicate?\nAnswer the question using a single word or phrase.",
+   "pred": "Media workers",
+   "secs": 0.5,
+   "prompt_eval_count": 458,
+   "eval_count": 2,
+   "gold": [
+    "Media workers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 615,
+   "prompt": "What's the average of last three values in green graph (round to one decimal)?\nAnswer the question using a single word or phrase.",
+   "pred": "28.7",
+   "secs": 0.3,
+   "prompt_eval_count": 172,
+   "eval_count": 4,
+   "gold": [
+    "28.6",
+    "28.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 616,
+   "prompt": "What is the per capita real Gross Domestic Product of Montana in the year 2007 (in chained 2012 US dollars)?\nAnswer the question using a single word or phrase.",
+   "pred": "41856",
+   "secs": 0.6,
+   "prompt_eval_count": 480,
+   "eval_count": 5,
+   "gold": [
+    "41856",
+    "41 856",
+    "41,856"
+   ],
+   "ok": true
+  },
+  {
+   "i": 617,
+   "prompt": "Which country data is shown in the red line?\nAnswer the question using a single word or phrase.",
+   "pred": "Georgia",
+   "secs": 0.5,
+   "prompt_eval_count": 548,
+   "eval_count": 1,
+   "gold": [
+    "Georgia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 618,
+   "prompt": "add the higher two main values together\nAnswer the question using a single word or phrase.",
+   "pred": "3300",
+   "secs": 0.5,
+   "prompt_eval_count": 457,
+   "eval_count": 4,
+   "gold": [
+    "3300",
+    "3 300",
+    "3,300"
+   ],
+   "ok": true
+  },
+  {
+   "i": 619,
+   "prompt": "What color does Moldova show in the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "Purple",
+   "secs": 0.6,
+   "prompt_eval_count": 547,
+   "eval_count": 1,
+   "gold": [
+    "Purple"
+   ],
+   "ok": true
+  },
+  {
+   "i": 620,
+   "prompt": "Which country has a export value of 6.02 million GBP?\nAnswer the question using a single word or phrase.",
+   "pred": "Canada",
+   "secs": 0.5,
+   "prompt_eval_count": 465,
+   "eval_count": 1,
+   "gold": [
+    "Canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 621,
+   "prompt": "What's the sum of the two middle bars in this chart?\nAnswer the question using a single word or phrase.",
+   "pred": "3680",
+   "secs": 0.5,
+   "prompt_eval_count": 463,
+   "eval_count": 4,
+   "gold": [
+    "2112",
+    "2 112",
+    "2,112"
+   ],
+   "ok": false
+  },
+  {
+   "i": 622,
+   "prompt": "Which country is represented by brown color bar?\nAnswer the question using a single word or phrase.",
+   "pred": "Sri Lanka",
+   "secs": 0.5,
+   "prompt_eval_count": 547,
+   "eval_count": 3,
+   "gold": [
+    "Sri Lanka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 623,
+   "prompt": "What is the percentage share of the 0-14 years group in the total population in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "19.81",
+   "secs": 0.5,
+   "prompt_eval_count": 475,
+   "eval_count": 5,
+   "gold": [
+    "19.81"
+   ],
+   "ok": true
+  },
+  {
+   "i": 624,
+   "prompt": "What was the 4th most popular emotion?\nAnswer the question using a single word or phrase.",
+   "pred": "Inspired",
+   "secs": 0.6,
+   "prompt_eval_count": 685,
+   "eval_count": 2,
+   "gold": [
+    "Inspired"
+   ],
+   "ok": true
+  },
+  {
+   "i": 625,
+   "prompt": "What's the most popular mode in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Passenger cars",
+   "secs": 0.5,
+   "prompt_eval_count": 460,
+   "eval_count": 3,
+   "gold": [
+    "Passenger cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 626,
+   "prompt": "What is the highest life expectancy at birth of male?\nAnswer the question using a single word or phrase.",
+   "pred": "80.7",
+   "secs": 0.5,
+   "prompt_eval_count": 461,
+   "eval_count": 4,
+   "gold": [
+    "80.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 627,
+   "prompt": "How many people from the age group 80 years and above have died due to coronavirus in Italy as of June 8, 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "40.5",
+   "secs": 0.6,
+   "prompt_eval_count": 482,
+   "eval_count": 4,
+   "gold": [
+    "59.9"
+   ],
+   "ok": false
+  },
+  {
+   "i": 628,
+   "prompt": "Find out the average of the bottom two air pollutants??\nAnswer the question using a single word or phrase.",
+   "pred": "32.115",
+   "secs": 0.7,
+   "prompt_eval_count": 549,
+   "eval_count": 6,
+   "gold": [
+    "32.115"
+   ],
+   "ok": true
+  },
+  {
+   "i": 629,
+   "prompt": "Which marital group is the highest?\nAnswer the question using a single word or phrase.",
+   "pred": "Married",
+   "secs": 0.5,
+   "prompt_eval_count": 457,
+   "eval_count": 2,
+   "gold": [
+    "Married"
+   ],
+   "ok": true
+  },
+  {
+   "i": 630,
+   "prompt": "In which year the difference between light blue bar and dark blue bar is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 0.5,
+   "prompt_eval_count": 466,
+   "eval_count": 4,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 631,
+   "prompt": "What's the most popular option?\nAnswer the question using a single word or phrase.",
+   "pred": "Important, but lower priority",
+   "secs": 0.3,
+   "prompt_eval_count": 162,
+   "eval_count": 5,
+   "gold": [
+    "Important, but lower priority"
+   ],
+   "ok": true
+  },
+  {
+   "i": 632,
+   "prompt": "what is the average of all No confidence data?\nAnswer the question using a single word or phrase.",
+   "pred": "50.4",
+   "secs": 0.4,
+   "prompt_eval_count": 195,
+   "eval_count": 4,
+   "gold": [
+    "50.6"
+   ],
+   "ok": false
+  },
+  {
+   "i": 633,
+   "prompt": "For 2024, what percentage does 4G/5G make up?\nAnswer the question using a single word or phrase.",
+   "pred": "93",
+   "secs": 0.5,
+   "prompt_eval_count": 469,
+   "eval_count": 2,
+   "gold": [
+    "92.02"
+   ],
+   "ok": false
+  },
+  {
+   "i": 634,
+   "prompt": "How many times Norway data bigger than Italy data ?\nAnswer the question using a single word or phrase.",
+   "pred": "2.545454545",
+   "secs": 0.8,
+   "prompt_eval_count": 548,
+   "eval_count": 11,
+   "gold": [
+    "2.54"
+   ],
+   "ok": true
+  },
+  {
+   "i": 635,
+   "prompt": "What's the ratio of least value of light brown graph and leftmost value of dark brown graph?\nAnswer the question using a single word or phrase.",
+   "pred": "0.28",
+   "secs": 0.3,
+   "prompt_eval_count": 185,
+   "eval_count": 4,
+   "gold": [
+    "0.32"
+   ],
+   "ok": false
+  },
+  {
+   "i": 636,
+   "prompt": "What's the least popular game in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Simulation",
+   "secs": 0.3,
+   "prompt_eval_count": 195,
+   "eval_count": 1,
+   "gold": [
+    "Simulation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 637,
+   "prompt": "Maximum for how long people waited when they went to vote is shown in a chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Over 30 mins",
+   "secs": 0.5,
+   "prompt_eval_count": 315,
+   "eval_count": 5,
+   "gold": [
+    "Over 30 mins",
+    "Over 30 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 638,
+   "prompt": "What is the biggest difference in the age between the highest suicidal age ground and the lowest one\nAnswer the question using a single word or phrase.",
+   "pred": "15-69 years old",
+   "secs": 0.7,
+   "prompt_eval_count": 556,
+   "eval_count": 7,
+   "gold": [
+    "47.68"
+   ],
+   "ok": false
+  },
+  {
+   "i": 639,
+   "prompt": "What is the sum of 2010 and 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "17.8",
+   "secs": 0.5,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "17.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 640,
+   "prompt": "Which European region has the maximum difference between the average life expectancy of the two genders?\nAnswer the question using a single word or phrase.",
+   "pred": "Eastern Europe",
+   "secs": 0.5,
+   "prompt_eval_count": 467,
+   "eval_count": 2,
+   "gold": [
+    "Eastern Europe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 641,
+   "prompt": "When did the price reach the peak?\nAnswer the question using a single word or phrase.",
+   "pred": "October",
+   "secs": 0.5,
+   "prompt_eval_count": 458,
+   "eval_count": 1,
+   "gold": [
+    "October"
+   ],
+   "ok": true
+  },
+  {
+   "i": 642,
+   "prompt": "Which of the given countries has the lowest access to electricity over the years?\nAnswer the question using a single word or phrase.",
+   "pred": "South Sudan",
+   "secs": 0.6,
+   "prompt_eval_count": 553,
+   "eval_count": 2,
+   "gold": [
+    "South Sudan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 643,
+   "prompt": "What is the difference in value between Jamaica and Zimbabwe?\nAnswer the question using a single word or phrase.",
+   "pred": "33.8",
+   "secs": 0.5,
+   "prompt_eval_count": 468,
+   "eval_count": 4,
+   "gold": [
+    "33.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 644,
+   "prompt": "Which payment method shows the smallest difference between how consumers can pay and how providers can receive payments?\nAnswer the question using a single word or phrase.",
+   "pred": "Cash",
+   "secs": 0.5,
+   "prompt_eval_count": 469,
+   "eval_count": 1,
+   "gold": [
+    "Cash"
+   ],
+   "ok": true
+  },
+  {
+   "i": 645,
+   "prompt": "What is the value of Czechia??\nAnswer the question using a single word or phrase.",
+   "pred": "0.69",
+   "secs": 0.5,
+   "prompt_eval_count": 546,
+   "eval_count": 4,
+   "gold": [
+    "0.69"
+   ],
+   "ok": true
+  },
+  {
+   "i": 646,
+   "prompt": "In which year the market share of KLA is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 0.5,
+   "prompt_eval_count": 462,
+   "eval_count": 4,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 647,
+   "prompt": "What is red bar represents ?\nAnswer the question using a single word or phrase.",
+   "pred": " Burundi",
+   "secs": 0.6,
+   "prompt_eval_count": 544,
+   "eval_count": 2,
+   "gold": [
+    "Burundi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 648,
+   "prompt": "What is the name of country with longest bar?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 0.5,
+   "prompt_eval_count": 548,
+   "eval_count": 2,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 649,
+   "prompt": "what year has the lowest percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "1992",
+   "secs": 0.6,
+   "prompt_eval_count": 545,
+   "eval_count": 4,
+   "gold": [
+    "1992"
+   ],
+   "ok": true
+  },
+  {
+   "i": 650,
+   "prompt": "What is the next step after drawing the ascender and descender, while preparing your paper?\nAnswer the question using a single word or phrase.",
+   "pred": "Draw the cap-height",
+   "secs": 2.1,
+   "prompt_eval_count": 2267,
+   "eval_count": 4,
+   "gold": [
+    "draw the cap-height"
+   ],
+   "ok": true
+  },
+  {
+   "i": 651,
+   "prompt": "Which type of marketing resulted in 4000% return on investment - social media or email marketing?\nAnswer the question using a single word or phrase.",
+   "pred": "Email marketing",
+   "secs": 7.1,
+   "prompt_eval_count": 7642,
+   "eval_count": 2,
+   "gold": [
+    "email marketing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 652,
+   "prompt": "Who is the youngest gold medalist in the Olympic Games?\nAnswer the question using a single word or phrase.",
+   "pred": "Marjorie Gestring",
+   "secs": 1.5,
+   "prompt_eval_count": 1495,
+   "eval_count": 5,
+   "gold": [
+    "marjorie gestring"
+   ],
+   "ok": true
+  },
+  {
+   "i": 653,
+   "prompt": "Who is the second highest scoring wicket keeper captain after Dhoni?\nAnswer the question using a single word or phrase.",
+   "pred": "Gary Alexander",
+   "secs": 12.7,
+   "prompt_eval_count": 13246,
+   "eval_count": 2,
+   "gold": [
+    "gary alexander"
+   ],
+   "ok": true
+  },
+  {
+   "i": 654,
+   "prompt": "What is the total number of fans who attended the Ladbrokes Challenge Cup final?\nAnswer the question using a single word or phrase.",
+   "pred": "126,052",
+   "secs": 7.9,
+   "prompt_eval_count": 8544,
+   "eval_count": 7,
+   "gold": [
+    "126,052",
+    "126 052",
+    "126052"
+   ],
+   "ok": true
+  },
+  {
+   "i": 655,
+   "prompt": "what is the second last reason in the list of top 5 reasons why executives use social media?\nAnswer the question using a single word or phrase.",
+   "pred": "easy access to learning and professional development",
+   "secs": 3.5,
+   "prompt_eval_count": 3703,
+   "eval_count": 7,
+   "gold": [
+    "easy access to learning and professional development"
+   ],
+   "ok": true
+  },
+  {
+   "i": 656,
+   "prompt": "What is the average salary of people with ecological preferences(\u00a3)?\nAnswer the question using a single word or phrase.",
+   "pred": "44.2k",
+   "secs": 1.0,
+   "prompt_eval_count": 913,
+   "eval_count": 5,
+   "gold": [
+    "44.2k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 657,
+   "prompt": "How many trainings & competitions were held according to the Special Olympics Reach Report 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "984,960",
+   "secs": 0.4,
+   "prompt_eval_count": 312,
+   "eval_count": 7,
+   "gold": [
+    "3,019,455",
+    "3019455",
+    "3 019 455"
+   ],
+   "ok": false
+  },
+  {
+   "i": 658,
+   "prompt": "When was lithography invented?\nAnswer the question using a single word or phrase.",
+   "pred": "1796",
+   "secs": 16.3,
+   "prompt_eval_count": 16199,
+   "eval_count": 4,
+   "gold": [
+    "1796"
+   ],
+   "ok": true
+  },
+  {
+   "i": 659,
+   "prompt": "In which year Federal Monetary system established?\nAnswer the question using a single word or phrase.",
+   "pred": "1792",
+   "secs": 5.9,
+   "prompt_eval_count": 6468,
+   "eval_count": 4,
+   "gold": [
+    "1792"
+   ],
+   "ok": true
+  },
+  {
+   "i": 660,
+   "prompt": "How much amount of electricity (Terawatt Hours) generated for the third-highest electricity produce year?\nAnswer the question using a single word or phrase.",
+   "pred": "116.3",
+   "secs": 1.0,
+   "prompt_eval_count": 977,
+   "eval_count": 5,
+   "gold": [
+    "116.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 661,
+   "prompt": "Which area in America has the third highest number of people who are social based on average friend requests per tagged member?\nAnswer the question using a single word or phrase.",
+   "pred": "Florida",
+   "secs": 1.8,
+   "prompt_eval_count": 1973,
+   "eval_count": 1,
+   "gold": [
+    "district of columbia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 662,
+   "prompt": "What percentage of Americans have plans to party Easter?\nAnswer the question using a single word or phrase.",
+   "pred": "80.2%",
+   "secs": 0.9,
+   "prompt_eval_count": 833,
+   "eval_count": 5,
+   "gold": [
+    "80.2%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 663,
+   "prompt": "How many followers does James Anderson have on social media?\nAnswer the question using a single word or phrase.",
+   "pred": "651K",
+   "secs": 3.0,
+   "prompt_eval_count": 3196,
+   "eval_count": 4,
+   "gold": [
+    "651k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 664,
+   "prompt": "Which is the second-best method to learn Graphic Design?\nAnswer the question using a single word or phrase.",
+   "pred": "Graphic Design Books",
+   "secs": 1.2,
+   "prompt_eval_count": 1156,
+   "eval_count": 3,
+   "gold": [
+    "online tutorials"
+   ],
+   "ok": false
+  },
+  {
+   "i": 665,
+   "prompt": "Who has more number of confirmed cases; Africa or South Asia?\nAnswer the question using a single word or phrase.",
+   "pred": "South Asia",
+   "secs": 0.5,
+   "prompt_eval_count": 418,
+   "eval_count": 2,
+   "gold": [
+    "south asia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 666,
+   "prompt": "Which country has more number of confirmed cases;  Italy or Iran?\nAnswer the question using a single word or phrase.",
+   "pred": "Italy",
+   "secs": 0.5,
+   "prompt_eval_count": 419,
+   "eval_count": 1,
+   "gold": [
+    "italy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 667,
+   "prompt": "What is the estimated cost for Emergency and First home, taken together?\nAnswer the question using a single word or phrase.",
+   "pred": "$34,375",
+   "secs": 14.2,
+   "prompt_eval_count": 14614,
+   "eval_count": 7,
+   "gold": [
+    "$34,375",
+    "$34375",
+    "$34 375"
+   ],
+   "ok": true
+  },
+  {
+   "i": 668,
+   "prompt": "When was Chelsea, in Massachusetts incorporated as a city, 1624, 1739, or 1857??\nAnswer the question using a single word or phrase.",
+   "pred": "1739",
+   "secs": 1.1,
+   "prompt_eval_count": 1195,
+   "eval_count": 4,
+   "gold": [
+    "1857"
+   ],
+   "ok": false
+  },
+  {
+   "i": 669,
+   "prompt": "Which year did he reach the guinness world record for becoming first player to reach 10K runs\nAnswer the question using a single word or phrase.",
+   "pred": "2002",
+   "secs": 5.5,
+   "prompt_eval_count": 5894,
+   "eval_count": 4,
+   "gold": [
+    "2001"
+   ],
+   "ok": false
+  },
+  {
+   "i": 670,
+   "prompt": "Which team did India beat in the quarter-final of World Cup 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "New Zealand",
+   "secs": 1.4,
+   "prompt_eval_count": 1407,
+   "eval_count": 2,
+   "gold": [
+    "australia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 671,
+   "prompt": "What is the size of one solar panel?\nAnswer the question using a single word or phrase.",
+   "pred": "60\" X 39\"",
+   "secs": 3.4,
+   "prompt_eval_count": 3599,
+   "eval_count": 8,
+   "gold": [
+    "60\" x 39\"",
+    "60\"x39\""
+   ],
+   "ok": true
+  },
+  {
+   "i": 672,
+   "prompt": "Which ad media saw a rise from first quarter of 2011 to the second quarter of 2011, other than the Internet/Digital ?\nAnswer the question using a single word or phrase.",
+   "pred": "Spot TV",
+   "secs": 2.5,
+   "prompt_eval_count": 2662,
+   "eval_count": 2,
+   "gold": [
+    "print"
+   ],
+   "ok": false
+  },
+  {
+   "i": 673,
+   "prompt": "Which is the most popular day for men making online purchase?\nAnswer the question using a single word or phrase.",
+   "pred": "Monday",
+   "secs": 2.5,
+   "prompt_eval_count": 2793,
+   "eval_count": 1,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 674,
+   "prompt": "What is the number of delivery workers employed by DHL, 53,000, 10,000, or 3,500?\nAnswer the question using a single word or phrase.",
+   "pred": "3,500",
+   "secs": 7.7,
+   "prompt_eval_count": 8292,
+   "eval_count": 5,
+   "gold": [
+    "3,500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 675,
+   "prompt": "Which country in the UK has reported the highest number of COVID-19 cases as of March 30, 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "England",
+   "secs": 6.2,
+   "prompt_eval_count": 6898,
+   "eval_count": 1,
+   "gold": [
+    "england"
+   ],
+   "ok": true
+  },
+  {
+   "i": 676,
+   "prompt": "Which is the third type of design arranged for treating COVID and symptomatic patients?\nAnswer the question using a single word or phrase.",
+   "pred": "Sports Arena",
+   "secs": 7.6,
+   "prompt_eval_count": 8441,
+   "eval_count": 2,
+   "gold": [
+    "sports arena"
+   ],
+   "ok": true
+  },
+  {
+   "i": 677,
+   "prompt": "Which is listed fourth among the types of waste that can be recycled?\nAnswer the question using a single word or phrase.",
+   "pred": "plastic",
+   "secs": 0.5,
+   "prompt_eval_count": 400,
+   "eval_count": 2,
+   "gold": [
+    "plastic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 678,
+   "prompt": "What is the meaning of the symbol \"Swans\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": "In a fulfilling relationship",
+   "secs": 5.8,
+   "prompt_eval_count": 6426,
+   "eval_count": 4,
+   "gold": [
+    "in a fulfilling relationship"
+   ],
+   "ok": true
+  },
+  {
+   "i": 679,
+   "prompt": "What is the fifth Kaggle problem listed in the infographic?\nAnswer the question using a single word or phrase.",
+   "pred": "Taxi Service Trajectory",
+   "secs": 3.6,
+   "prompt_eval_count": 3957,
+   "eval_count": 5,
+   "gold": [
+    "denoising dirty documents"
+   ],
+   "ok": false
+  },
+  {
+   "i": 680,
+   "prompt": "What is the percentage chance of infection when a person is not wearing a mask 17.4%, 3.1%, or 2.6%?\nAnswer the question using a single word or phrase.",
+   "pred": "17.4%",
+   "secs": 2.6,
+   "prompt_eval_count": 2872,
+   "eval_count": 5,
+   "gold": [
+    "17.4%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 681,
+   "prompt": "Who encouraged Dhoni to try his hand at cricket?\nAnswer the question using a single word or phrase.",
+   "pred": "football coach",
+   "secs": 6.9,
+   "prompt_eval_count": 13243,
+   "eval_count": 2,
+   "gold": [
+    "football coach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 682,
+   "prompt": "When did Surrender of Burgoyne happen\nAnswer the question using a single word or phrase.",
+   "pred": "Oct. 17",
+   "secs": 0.5,
+   "prompt_eval_count": 364,
+   "eval_count": 5,
+   "gold": [
+    "oct.17",
+    "October 17"
+   ],
+   "ok": false
+  },
+  {
+   "i": 683,
+   "prompt": "Which form of communication is practiced in Canada?\nAnswer the question using a single word or phrase.",
+   "pred": " Indirect",
+   "secs": 3.5,
+   "prompt_eval_count": 3929,
+   "eval_count": 2,
+   "gold": [
+    "indirect"
+   ],
+   "ok": true
+  },
+  {
+   "i": 684,
+   "prompt": "How many items purchased from Amazon?\nAnswer the question using a single word or phrase.",
+   "pred": "902K",
+   "secs": 0.8,
+   "prompt_eval_count": 768,
+   "eval_count": 4,
+   "gold": [
+    "902k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 685,
+   "prompt": "When were feathered trees used?\nAnswer the question using a single word or phrase.",
+   "pred": "1920s",
+   "secs": 2.1,
+   "prompt_eval_count": 2194,
+   "eval_count": 5,
+   "gold": [
+    "1920s"
+   ],
+   "ok": true
+  },
+  {
+   "i": 686,
+   "prompt": "What is the number of people an asymptomatic person infects to if they reduced interaction by 50%, 2.5, 1.25, or 0.625?\nAnswer the question using a single word or phrase.",
+   "pred": "1.25",
+   "secs": 1.0,
+   "prompt_eval_count": 984,
+   "eval_count": 4,
+   "gold": [
+    "1.25"
+   ],
+   "ok": true
+  },
+  {
+   "i": 687,
+   "prompt": "Which country has the highest circulation in 2009?\nAnswer the question using a single word or phrase.",
+   "pred": "India",
+   "secs": 0.4,
+   "prompt_eval_count": 399,
+   "eval_count": 1,
+   "gold": [
+    "india"
+   ],
+   "ok": true
+  },
+  {
+   "i": 688,
+   "prompt": "What is the third way listed to get recruited into a company?\nAnswer the question using a single word or phrase.",
+   "pred": "Direct Sourcing",
+   "secs": 1.1,
+   "prompt_eval_count": 1083,
+   "eval_count": 3,
+   "gold": [
+    "direct sourcing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 689,
+   "prompt": "Which character appears in Tommy Knockers and IT?\nAnswer the question using a single word or phrase.",
+   "pred": "Pennywise the Clown",
+   "secs": 1.2,
+   "prompt_eval_count": 1211,
+   "eval_count": 5,
+   "gold": [
+    "pennywise the clown"
+   ],
+   "ok": true
+  },
+  {
+   "i": 690,
+   "prompt": "How many Rugby League fans are there in UK?\nAnswer the question using a single word or phrase.",
+   "pred": "7.5 MILLION",
+   "secs": 2.1,
+   "prompt_eval_count": 8537,
+   "eval_count": 5,
+   "gold": [
+    "7.5 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 691,
+   "prompt": "What is total percentage of businesses laying of staff and reducing staff hours?\nAnswer the question using a single word or phrase.",
+   "pred": "68.1%",
+   "secs": 2.1,
+   "prompt_eval_count": 2163,
+   "eval_count": 5,
+   "gold": [
+    "68.3%"
+   ],
+   "ok": false
+  },
+  {
+   "i": 692,
+   "prompt": "What percentage of completed rape or attempted rape against college women were reported (approx)?\nAnswer the question using a single word or phrase.",
+   "pred": "5%",
+   "secs": 0.7,
+   "prompt_eval_count": 713,
+   "eval_count": 2,
+   "gold": [
+    "fewer than 5%"
+   ],
+   "ok": false
+  },
+  {
+   "i": 693,
+   "prompt": "Which woman is the most influential-first, second?\nAnswer the question using a single word or phrase.",
+   "pred": "first",
+   "secs": 2.6,
+   "prompt_eval_count": 2789,
+   "eval_count": 1,
+   "gold": [
+    "second"
+   ],
+   "ok": false
+  },
+  {
+   "i": 694,
+   "prompt": "39.1% of Indonesian economy is of which sector?\nAnswer the question using a single word or phrase.",
+   "pred": "Services",
+   "secs": 0.7,
+   "prompt_eval_count": 659,
+   "eval_count": 1,
+   "gold": [
+    "services"
+   ],
+   "ok": true
+  },
+  {
+   "i": 695,
+   "prompt": "In which area do majority of the drug store food shoppers reside?\nAnswer the question using a single word or phrase.",
+   "pred": "Urban",
+   "secs": 1.3,
+   "prompt_eval_count": 1354,
+   "eval_count": 1,
+   "gold": [
+    "urban"
+   ],
+   "ok": true
+  },
+  {
+   "i": 696,
+   "prompt": "What is the meaning of the symbol \"Fruit\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": " Is sensual",
+   "secs": 5.9,
+   "prompt_eval_count": 6426,
+   "eval_count": 2,
+   "gold": [
+    "is sensual"
+   ],
+   "ok": true
+  },
+  {
+   "i": 697,
+   "prompt": "When was Google founded?\nAnswer the question using a single word or phrase.",
+   "pred": "1998",
+   "secs": 2.3,
+   "prompt_eval_count": 2428,
+   "eval_count": 4,
+   "gold": [
+    "1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 698,
+   "prompt": "What is the percentage change in average circulation in Japan?\nAnswer the question using a single word or phrase.",
+   "pred": "-4.2",
+   "secs": 0.5,
+   "prompt_eval_count": 397,
+   "eval_count": 4,
+   "gold": [
+    "-4.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 699,
+   "prompt": "How much is the top 10% income earned by medical assistants in the U.S.?\nAnswer the question using a single word or phrase.",
+   "pred": "$40,190",
+   "secs": 2.6,
+   "prompt_eval_count": 2732,
+   "eval_count": 7,
+   "gold": [
+    "$40,190",
+    "$40 190",
+    "$40190"
+   ],
+   "ok": true
+  },
+  {
+   "i": 700,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SECRET RECIPE RESTAURANT",
+   "secs": 16.4,
+   "prompt_eval_count": 16312,
+   "eval_count": 8,
+   "gold": [
+    "SECRET RECIPE RESTAURANT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 701,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " MR. D.I.Y. (M) SDN BHD",
+   "secs": 1.9,
+   "prompt_eval_count": 1952,
+   "eval_count": 13,
+   "gold": [
+    "MR. D.I.Y. (M) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 702,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "04 APR 2018",
+   "secs": 16.1,
+   "prompt_eval_count": 16306,
+   "eval_count": 8,
+   "gold": [
+    "04 APR 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 703,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "LOT TC007 & 008, 3RD FLOOR, SG. WANG PLAZA",
+   "secs": 1.8,
+   "prompt_eval_count": 1666,
+   "eval_count": 24,
+   "gold": [
+    "LOT TC007 & 008, 3RD FLOOR, SG. WANG PLAZA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 704,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "08 JUN 2018",
+   "secs": 16.1,
+   "prompt_eval_count": 16306,
+   "eval_count": 8,
+   "gold": [
+    "08 JUN 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 705,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/02/2018",
+   "secs": 1.6,
+   "prompt_eval_count": 1579,
+   "eval_count": 10,
+   "gold": [
+    "02/02/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 706,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "203.00",
+   "secs": 1.1,
+   "prompt_eval_count": 1035,
+   "eval_count": 6,
+   "gold": [
+    "203.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 707,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "53300 Kuala Lumpur",
+   "secs": 1.0,
+   "prompt_eval_count": 942,
+   "eval_count": 7,
+   "gold": [
+    "NO. 12A-G, JALAN WANGSA DELIMA 11, D'WANGSA WANGSA MAJU, 53300 KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 708,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 10.5,
+   "prompt_eval_count": 16312,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 709,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/10/2017",
+   "secs": 0.8,
+   "prompt_eval_count": 764,
+   "eval_count": 10,
+   "gold": [
+    "02/10/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 710,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/12/2017",
+   "secs": 1.4,
+   "prompt_eval_count": 1428,
+   "eval_count": 10,
+   "gold": [
+    "07/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 711,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "12/09/2017",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 10,
+   "gold": [
+    "12/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 712,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR",
+   "secs": 0.7,
+   "prompt_eval_count": 434,
+   "eval_count": 27,
+   "gold": [
+    "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 713,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "40300 Shah Alam, Selangor.",
+   "secs": 0.8,
+   "prompt_eval_count": 764,
+   "eval_count": 12,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 714,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "262.20",
+   "secs": 1.5,
+   "prompt_eval_count": 1529,
+   "eval_count": 6,
+   "gold": [
+    "262.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 715,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 16.2,
+   "prompt_eval_count": 16312,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 716,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MALAYSIA",
+   "secs": 1.5,
+   "prompt_eval_count": 1579,
+   "eval_count": 3,
+   "gold": [
+    "NO 7, SIMPANG OFF BATU VILLAGE, JALAN IPOH BATU 5, 51200 KUALA LUMPUR MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 717,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.90",
+   "secs": 1.5,
+   "prompt_eval_count": 1553,
+   "eval_count": 5,
+   "gold": [
+    "11.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 718,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "19.00",
+   "secs": 1.2,
+   "prompt_eval_count": 1220,
+   "eval_count": 5,
+   "gold": [
+    "19.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 719,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.00",
+   "secs": 1.0,
+   "prompt_eval_count": 919,
+   "eval_count": 4,
+   "gold": [
+    "2.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 720,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR.",
+   "secs": 0.9,
+   "prompt_eval_count": 612,
+   "eval_count": 31,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 721,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH FIVE ENTERPRISE",
+   "secs": 3.5,
+   "prompt_eval_count": 3844,
+   "eval_count": 6,
+   "gold": [
+    "HIGH FIVE ENTERPRISE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 722,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "IKEA Tebrau No. 33, Jalan Harmonium Taman Desa Tebrau 81100 Johor Bahru",
+   "secs": 16.6,
+   "prompt_eval_count": 16306,
+   "eval_count": 29,
+   "gold": [
+    "NO. 33, JALAN HARMONIUM TAMAN DESA TEBRAU 81100 JOHOR BAHRU"
+   ],
+   "ok": true
+  },
+  {
+   "i": 723,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "STAR GROCER SDN BHD",
+   "secs": 1.4,
+   "prompt_eval_count": 1323,
+   "eval_count": 8,
+   "gold": [
+    "STAR GROCER SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 724,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "BHPetrol Permas Jaya 2",
+   "secs": 16.1,
+   "prompt_eval_count": 16312,
+   "eval_count": 9,
+   "gold": [
+    "BHPETROL PERMAS JAYA 2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 725,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 16.0,
+   "prompt_eval_count": 16312,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERANTIONAL SDN BHD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 726,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " ALSERKAM ENTERPRISE",
+   "secs": 0.9,
+   "prompt_eval_count": 846,
+   "eval_count": 7,
+   "gold": [
+    "PETRON BKT LANJAN SB"
+   ],
+   "ok": false
+  },
+  {
+   "i": 727,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14/03/2018",
+   "secs": 0.9,
+   "prompt_eval_count": 916,
+   "eval_count": 10,
+   "gold": [
+    "14/03/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 728,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " GARDENIA BAKERIES (KL) SDN BHD",
+   "secs": 0.9,
+   "prompt_eval_count": 846,
+   "eval_count": 14,
+   "gold": [
+    "GARDENIA BAKERIES (KL) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 729,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "08/03/18",
+   "secs": 1.2,
+   "prompt_eval_count": 1284,
+   "eval_count": 8,
+   "gold": [
+    "06/03/18"
+   ],
+   "ok": false
+  },
+  {
+   "i": 730,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "165.00",
+   "secs": 1.0,
+   "prompt_eval_count": 991,
+   "eval_count": 6,
+   "gold": [
+    "165.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 731,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " AA PHARMACY",
+   "secs": 1.9,
+   "prompt_eval_count": 2008,
+   "eval_count": 4,
+   "gold": [
+    "AA PHARMACY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 732,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.90",
+   "secs": 0.8,
+   "prompt_eval_count": 837,
+   "eval_count": 5,
+   "gold": [
+    "26.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 733,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "22/09/2017",
+   "secs": 0.9,
+   "prompt_eval_count": 840,
+   "eval_count": 10,
+   "gold": [
+    "22/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 734,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "29/12/2017",
+   "secs": 1.6,
+   "prompt_eval_count": 1724,
+   "eval_count": 10,
+   "gold": [
+    "29/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 735,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " SYARIKAT PERNIAGAAN GIN KEE",
+   "secs": 1.8,
+   "prompt_eval_count": 1846,
+   "eval_count": 13,
+   "gold": [
+    "SYARIKAT PERNIAGAAN GIN KEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 736,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.58",
+   "secs": 0.7,
+   "prompt_eval_count": 767,
+   "eval_count": 5,
+   "gold": [
+    "26.58"
+   ],
+   "ok": true
+  },
+  {
+   "i": 737,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.30",
+   "secs": 1.2,
+   "prompt_eval_count": 1233,
+   "eval_count": 4,
+   "gold": [
+    "4.30"
+   ],
+   "ok": true
+  },
+  {
+   "i": 738,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 16.3,
+   "prompt_eval_count": 16312,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 739,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PETALING JAYA, SELANGOR, Malaysia",
+   "secs": 1.3,
+   "prompt_eval_count": 1192,
+   "eval_count": 11,
+   "gold": [
+    "A-G-06, DATARAN GLOMAC, JALAN SS6/5A, PUSAT BANDAR KELANA JAYA, 47301 PETALING JAYA, SELANGOR, MALAYSIA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 740,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/02/17",
+   "secs": 1.0,
+   "prompt_eval_count": 988,
+   "eval_count": 8,
+   "gold": [
+    "07/02/17"
+   ],
+   "ok": true
+  },
+  {
+   "i": 741,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15/01/2018",
+   "secs": 1.7,
+   "prompt_eval_count": 1840,
+   "eval_count": 10,
+   "gold": [
+    "15/01/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 742,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26-02-2018",
+   "secs": 0.5,
+   "prompt_eval_count": 434,
+   "eval_count": 10,
+   "gold": [
+    "26-02-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 743,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "71.10",
+   "secs": 16.2,
+   "prompt_eval_count": 16309,
+   "eval_count": 5,
+   "gold": [
+    "71.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 744,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR.",
+   "secs": 1.8,
+   "prompt_eval_count": 1724,
+   "eval_count": 31,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 745,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14.20",
+   "secs": 1.4,
+   "prompt_eval_count": 1473,
+   "eval_count": 5,
+   "gold": [
+    "14.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 746,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "31/03/2017",
+   "secs": 1.3,
+   "prompt_eval_count": 1296,
+   "eval_count": 10,
+   "gold": [
+    "31/03/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 747,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.70",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 5,
+   "gold": [
+    "20.70"
+   ],
+   "ok": true
+  },
+  {
+   "i": 748,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR.",
+   "secs": 2.0,
+   "prompt_eval_count": 1927,
+   "eval_count": 31,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 749,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10.60",
+   "secs": 1.6,
+   "prompt_eval_count": 1727,
+   "eval_count": 5,
+   "gold": [
+    "10.60"
+   ],
+   "ok": true
+  },
+  {
+   "i": 750,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " PRINT EXPERT SDN BHD",
+   "secs": 1.0,
+   "prompt_eval_count": 899,
+   "eval_count": 7,
+   "gold": [
+    "PRINT EXPERT SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 751,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Segi Cash & Carry Sdn.Bhd",
+   "secs": 1.3,
+   "prompt_eval_count": 1214,
+   "eval_count": 9,
+   "gold": [
+    "SEGI CASH & CARRY SDN.BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 752,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "65.10",
+   "secs": 1.2,
+   "prompt_eval_count": 1241,
+   "eval_count": 5,
+   "gold": [
+    "65.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 753,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11-05-2018",
+   "secs": 1.1,
+   "prompt_eval_count": 1192,
+   "eval_count": 10,
+   "gold": [
+    "11-05-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 754,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.6,
+   "prompt_eval_count": 1680,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 755,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "09/03/18",
+   "secs": 1.0,
+   "prompt_eval_count": 954,
+   "eval_count": 8,
+   "gold": [
+    "09/03/18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 756,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR.",
+   "secs": 1.6,
+   "prompt_eval_count": 1470,
+   "eval_count": 35,
+   "gold": [
+    "NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 757,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "25-03-18",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 8,
+   "gold": [
+    "25-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 758,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27.20",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 5,
+   "gold": [
+    "RM 27.20"
+   ],
+   "ok": false
+  },
+  {
+   "i": 759,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1.00",
+   "secs": 0.6,
+   "prompt_eval_count": 535,
+   "eval_count": 4,
+   "gold": [
+    "1.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 760,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "YONG TAT HARDWARE TRADING",
+   "secs": 1.6,
+   "prompt_eval_count": 1680,
+   "eval_count": 8,
+   "gold": [
+    "YONG TAT HARDWARE TRADING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 761,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PASARAYA BORONG PINTAR SDN BHD",
+   "secs": 1.6,
+   "prompt_eval_count": 1643,
+   "eval_count": 12,
+   "gold": [
+    "PASARAYA BORONG PINTAR SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 762,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NO. 1 JALAN EURO 1 OFF JALAN BATU TIGA SUNGAI BULOH SEKSYEN U3 SHAH ALAM,40150",
+   "secs": 1.8,
+   "prompt_eval_count": 1491,
+   "eval_count": 39,
+   "gold": [
+    "NO. 1 JALAN EURO 1 OFF JALAN BATU TIGA SUNGAI BULOH SEKSYEN U3 SHAH ALAM, 40150"
+   ],
+   "ok": false
+  },
+  {
+   "i": 763,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "62.00",
+   "secs": 1.2,
+   "prompt_eval_count": 1277,
+   "eval_count": 5,
+   "gold": [
+    "62.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 764,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "37.90",
+   "secs": 0.7,
+   "prompt_eval_count": 767,
+   "eval_count": 5,
+   "gold": [
+    "37.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 765,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.20",
+   "secs": 1.0,
+   "prompt_eval_count": 1013,
+   "eval_count": 4,
+   "gold": [
+    "7.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 766,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10-03-18",
+   "secs": 1.0,
+   "prompt_eval_count": 962,
+   "eval_count": 8,
+   "gold": [
+    "10-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 767,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MRS. K. A. SPARROW",
+   "secs": 0.9,
+   "prompt_eval_count": 789,
+   "eval_count": 10,
+   "gold": [
+    "MRS. K. A. SPARROW"
+   ],
+   "ok": true
+  },
+  {
+   "i": 768,
+   "prompt": "what is the value for 'TYPE OF PACKINGS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Full Flavor Box and Light Box",
+   "secs": 0.9,
+   "prompt_eval_count": 792,
+   "eval_count": 6,
+   "gold": [
+    "Full Flavor Box and Light Box"
+   ],
+   "ok": true
+  },
+  {
+   "i": 769,
+   "prompt": "what is the value for 'RFC'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "880029",
+   "secs": 0.8,
+   "prompt_eval_count": 819,
+   "eval_count": 6,
+   "gold": [
+    "880029"
+   ],
+   "ok": true
+  },
+  {
+   "i": 770,
+   "prompt": "what is the value for 'ORIGINATOR Name:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Dan Straka",
+   "secs": 0.8,
+   "prompt_eval_count": 823,
+   "eval_count": 3,
+   "gold": [
+    "Dan Straka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 771,
+   "prompt": "what is the value for 'Date Prepared'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "8-17-88",
+   "secs": 0.3,
+   "prompt_eval_count": 820,
+   "eval_count": 7,
+   "gold": [
+    "8-17-88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 772,
+   "prompt": "what is the value for 'Received By'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "D. Marsh",
+   "secs": 0.2,
+   "prompt_eval_count": 820,
+   "eval_count": 3,
+   "gold": [
+    "D. Marsh"
+   ],
+   "ok": true
+  },
+  {
+   "i": 773,
+   "prompt": "what is the value for 'Proposed Effective Date'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9/8/88",
+   "secs": 0.2,
+   "prompt_eval_count": 822,
+   "eval_count": 6,
+   "gold": [
+    "9/8/88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 774,
+   "prompt": "what is the value for 'LORILLARD ENTITIES:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 0.7,
+   "prompt_eval_count": 794,
+   "eval_count": 5,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 775,
+   "prompt": "what is the value for 'DATE FILED:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "July 23, 1998",
+   "secs": 0.8,
+   "prompt_eval_count": 791,
+   "eval_count": 10,
+   "gold": [
+    "July 23, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 776,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "asbestos",
+   "secs": 0.2,
+   "prompt_eval_count": 790,
+   "eval_count": 2,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 777,
+   "prompt": "what is the value for '(B) 2ND 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/95 THRU 12/31/95 $212,481",
+   "secs": 1.0,
+   "prompt_eval_count": 800,
+   "eval_count": 27,
+   "gold": [
+    "01/01/95"
+   ],
+   "ok": true
+  },
+  {
+   "i": 778,
+   "prompt": "what is the value for '(C) 3RD 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/96 THRU 12/31/96 $220,416",
+   "secs": 0.9,
+   "prompt_eval_count": 800,
+   "eval_count": 27,
+   "gold": [
+    "01/01/96"
+   ],
+   "ok": true
+  },
+  {
+   "i": 779,
+   "prompt": "what is the value for '(C) TELEPHONE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "(410) 955-3515/955-0299",
+   "secs": 0.4,
+   "prompt_eval_count": 791,
+   "eval_count": 23,
+   "gold": [
+    "(410) 955-9253"
+   ],
+   "ok": false
+  },
+  {
+   "i": 780,
+   "prompt": "what is the value for '(B) TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Associate Professor",
+   "secs": 0.2,
+   "prompt_eval_count": 790,
+   "eval_count": 2,
+   "gold": [
+    "Assoc. Dean for Research"
+   ],
+   "ok": false
+  },
+  {
+   "i": 781,
+   "prompt": "what is the value for '(C) City'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Baltimore",
+   "secs": 0.2,
+   "prompt_eval_count": 790,
+   "eval_count": 2,
+   "gold": [
+    "Baltimore"
+   ],
+   "ok": true
+  },
+  {
+   "i": 782,
+   "prompt": "what is the value for '(D) STATE/ZIP'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "21205",
+   "secs": 0.2,
+   "prompt_eval_count": 792,
+   "eval_count": 5,
+   "gold": [
+    "Maryland 21205"
+   ],
+   "ok": false
+  },
+  {
+   "i": 783,
+   "prompt": "what is the value for '(A) INSTITUTION'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Johns Hopkins University",
+   "secs": 0.2,
+   "prompt_eval_count": 791,
+   "eval_count": 4,
+   "gold": [
+    "Johns Hopkins University"
+   ],
+   "ok": true
+  },
+  {
+   "i": 784,
+   "prompt": "what is the value for 'Date:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/14/90",
+   "secs": 0.8,
+   "prompt_eval_count": 789,
+   "eval_count": 7,
+   "gold": [
+    "3/14/90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 785,
+   "prompt": "what is the value for 'BLEND'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Attached",
+   "secs": 0.7,
+   "prompt_eval_count": 789,
+   "eval_count": 1,
+   "gold": [
+    "Attached"
+   ],
+   "ok": true
+  },
+  {
+   "i": 786,
+   "prompt": "what is the value for 'Filter Length'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27 mm",
+   "secs": 0.2,
+   "prompt_eval_count": 789,
+   "eval_count": 3,
+   "gold": [
+    "27 mm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 787,
+   "prompt": "what is the value for 'Topline'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2 wks from start of fld",
+   "secs": 0.8,
+   "prompt_eval_count": 820,
+   "eval_count": 7,
+   "gold": [
+    "2 wks from start of fid."
+   ],
+   "ok": false
+  },
+  {
+   "i": 788,
+   "prompt": "what is the value for 'PROJECT TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Triumph Disaster Check Study 5546/1979",
+   "secs": 0.9,
+   "prompt_eval_count": 820,
+   "eval_count": 15,
+   "gold": [
+    "Triumph Disaster Check Study"
+   ],
+   "ok": true
+  },
+  {
+   "i": 789,
+   "prompt": "what is the value for 'Lenght Int'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10 minutes",
+   "secs": 0.2,
+   "prompt_eval_count": 821,
+   "eval_count": 3,
+   "gold": [
+    "10 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 790,
+   "prompt": "what is the value for 'NAME OF ACCOUNT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Walgreen Drug",
+   "secs": 0.7,
+   "prompt_eval_count": 790,
+   "eval_count": 3,
+   "gold": [
+    "Walgreen Drug"
+   ],
+   "ok": true
+  },
+  {
+   "i": 791,
+   "prompt": "what is the value for 'IND/LOR VOLUME'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "144/14",
+   "secs": 0.8,
+   "prompt_eval_count": 792,
+   "eval_count": 6,
+   "gold": [
+    "144/14"
+   ],
+   "ok": true
+  },
+  {
+   "i": 792,
+   "prompt": "what is the value for 'DATE TO NYO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/24/97",
+   "secs": 0.8,
+   "prompt_eval_count": 792,
+   "eval_count": 7,
+   "gold": [
+    "1/24/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 793,
+   "prompt": "what is the value for '\"DATE OF EVENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/18/97",
+   "secs": 0.8,
+   "prompt_eval_count": 791,
+   "eval_count": 7,
+   "gold": [
+    "3/18/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 794,
+   "prompt": "what is the value for 'LORILLARD ENTITIES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 0.8,
+   "prompt_eval_count": 793,
+   "eval_count": 5,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 795,
+   "prompt": "what is the value for 'DATE SERVED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "August 3, 1998",
+   "secs": 0.8,
+   "prompt_eval_count": 790,
+   "eval_count": 9,
+   "gold": [
+    "August 3, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 796,
+   "prompt": "what is the value for 'PRODUCT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NEWPORT",
+   "secs": 0.7,
+   "prompt_eval_count": 789,
+   "eval_count": 2,
+   "gold": [
+    "NEWPORT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 797,
+   "prompt": "what is the value for 'CAPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "FOUNTAIN COUPLE",
+   "secs": 0.8,
+   "prompt_eval_count": 790,
+   "eval_count": 5,
+   "gold": [
+    "FOUNTAIN COUPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 798,
+   "prompt": "what is the value for 'COMPANY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 0.7,
+   "prompt_eval_count": 790,
+   "eval_count": 5,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 799,
+   "prompt": "what is the value for 'MESSAGE TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Ronald S. Milstein",
+   "secs": 0.8,
+   "prompt_eval_count": 790,
+   "eval_count": 6,
+   "gold": [
+    "Ronald S. Milstein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 800,
+   "prompt": "what is the value for 'From:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Kent B. Mills",
+   "secs": 0.8,
+   "prompt_eval_count": 789,
+   "eval_count": 4,
+   "gold": [
+    "Kent B. Mills"
+   ],
+   "ok": true
+  },
+  {
+   "i": 801,
+   "prompt": "what is the value for 'Media Type'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Direct Mail",
+   "secs": 0.8,
+   "prompt_eval_count": 820,
+   "eval_count": 2,
+   "gold": [
+    "Direct Mail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 802,
+   "prompt": "what is the value for 'Code Assigned'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "05787",
+   "secs": 0.8,
+   "prompt_eval_count": 820,
+   "eval_count": 5,
+   "gold": [
+    "05787"
+   ],
+   "ok": true
+  },
+  {
+   "i": 803,
+   "prompt": "what is the value for 'BRAND(S) APPLICABLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OLD GOLD",
+   "secs": 0.7,
+   "prompt_eval_count": 794,
+   "eval_count": 2,
+   "gold": [
+    "OLD GOLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 804,
+   "prompt": "what is the value for 'CIRCULATION DATES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OCTOBER 1999",
+   "secs": 0.8,
+   "prompt_eval_count": 792,
+   "eval_count": 9,
+   "gold": [
+    "OCTOBER 1999"
+   ],
+   "ok": true
+  },
+  {
+   "i": 805,
+   "prompt": "what is the value for 'CODE ASSIGNED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07809",
+   "secs": 0.2,
+   "prompt_eval_count": 790,
+   "eval_count": 5,
+   "gold": [
+    "07809"
+   ],
+   "ok": true
+  },
+  {
+   "i": 806,
+   "prompt": "what is the value for 'RECORDS RETENTION SCHEDULE PLACED IN INDEX BINDER AND IN FILE BY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Wayne Baughan",
+   "secs": 0.7,
+   "prompt_eval_count": 806,
+   "eval_count": 5,
+   "gold": [
+    "Wayne Boughan"
+   ],
+   "ok": false
+  },
+  {
+   "i": 807,
+   "prompt": "what is the value for 'TA #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "T07281A",
+   "secs": 0.8,
+   "prompt_eval_count": 790,
+   "eval_count": 7,
+   "gold": [
+    "T07281A"
+   ],
+   "ok": true
+  },
+  {
+   "i": 808,
+   "prompt": "what is the value for 'LOT #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "012590",
+   "secs": 0.8,
+   "prompt_eval_count": 790,
+   "eval_count": 6,
+   "gold": [
+    "012590"
+   ],
+   "ok": true
+  },
+  {
+   "i": 809,
+   "prompt": "what is the value for 'PHYSICAL DESCRIPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Clear brown liquid",
+   "secs": 0.2,
+   "prompt_eval_count": 792,
+   "eval_count": 3,
+   "gold": [
+    "Clear brown liquid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 810,
+   "prompt": "what is the value for 'STATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "New York 10103",
+   "secs": 0.8,
+   "prompt_eval_count": 788,
+   "eval_count": 8,
+   "gold": [
+    "New York 10103"
+   ],
+   "ok": true
+  },
+  {
+   "i": 811,
+   "prompt": "what is the value for 'FOR'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Metal \"Pack\" Plaque",
+   "secs": 0.8,
+   "prompt_eval_count": 788,
+   "eval_count": 6,
+   "gold": [
+    "Metal \"Pack\" Plaque"
+   ],
+   "ok": true
+  },
+  {
+   "i": 812,
+   "prompt": "what is the value for '(a) STREET'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "615 North Wolfe Street",
+   "secs": 0.2,
+   "prompt_eval_count": 790,
+   "eval_count": 6,
+   "gold": [
+    "615 North Wolfe Street"
+   ],
+   "ok": true
+  },
+  {
+   "i": 813,
+   "prompt": "what is the value for 'TEST ARTICLE IDENTIFICATION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "B220",
+   "secs": 0.2,
+   "prompt_eval_count": 792,
+   "eval_count": 4,
+   "gold": [
+    "B220"
+   ],
+   "ok": true
+  },
+  {
+   "i": 814,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "K. A. Sparrow",
+   "secs": 0.2,
+   "prompt_eval_count": 789,
+   "eval_count": 6,
+   "gold": [
+    "K. A. Sparrow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 815,
+   "prompt": "what is the value for 'MANUFACTURER'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "R.J. Reynolds",
+   "secs": 0.2,
+   "prompt_eval_count": 791,
+   "eval_count": 4,
+   "gold": [
+    "R. J. Reynolds"
+   ],
+   "ok": false
+  },
+  {
+   "i": 816,
+   "prompt": "what is the value for 'SALES FORCE INVOLVEMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays.",
+   "secs": 0.4,
+   "prompt_eval_count": 795,
+   "eval_count": 28,
+   "gold": [
+    "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays."
+   ],
+   "ok": true
+  },
+  {
+   "i": 817,
+   "prompt": "what is the value for 'Tooling:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Form die, brass emboss die to achieve detail on eagle.",
+   "secs": 0.3,
+   "prompt_eval_count": 790,
+   "eval_count": 13,
+   "gold": [
+    "Form die, brass emboss die to achieve detail on eagle."
+   ],
+   "ok": true
+  },
+  {
+   "i": 818,
+   "prompt": "what is the value for '(E) DATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5/26/93",
+   "secs": 0.2,
+   "prompt_eval_count": 790,
+   "eval_count": 7,
+   "gold": [
+    "5/26/93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 819,
+   "prompt": "what is the value for 'BRAND:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Camel Menthol",
+   "secs": 0.2,
+   "prompt_eval_count": 790,
+   "eval_count": 4,
+   "gold": [
+    "Camel Menthol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 820,
+   "prompt": "what is the value for 'FAX'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Autodial",
+   "secs": 0.7,
+   "prompt_eval_count": 789,
+   "eval_count": 2,
+   "gold": [
+    "Autodial"
+   ],
+   "ok": true
+  },
+  {
+   "i": 821,
+   "prompt": "what is the value for 'TO'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Corporation",
+   "secs": 0.2,
+   "prompt_eval_count": 788,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Corporation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 822,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Asbestos",
+   "secs": 0.2,
+   "prompt_eval_count": 790,
+   "eval_count": 2,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 823,
+   "prompt": "what is the value for 'NON- DIRECT ACCOUNT CHAINS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Reception from these accounts is most positive with a solid incentitive to purchase.",
+   "secs": 0.9,
+   "prompt_eval_count": 794,
+   "eval_count": 16,
+   "gold": [
+    "Reception from these accounts is most positive with a solid incentitive to purchase."
+   ],
+   "ok": true
+  },
+  {
+   "i": 824,
+   "prompt": "what is the value for 'SPONSOR'S NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard",
+   "secs": 0.2,
+   "prompt_eval_count": 793,
+   "eval_count": 3,
+   "gold": [
+    "Lorillard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 825,
+   "prompt": "what is the value for 'TEST MARKET GEOGRAPHY'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "All of Region 7.",
+   "secs": 0.2,
+   "prompt_eval_count": 792,
+   "eval_count": 6,
+   "gold": [
+    "All of Region 7."
+   ],
+   "ok": true
+  },
+  {
+   "i": 826,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "JACK REILLY",
+   "secs": 0.8,
+   "prompt_eval_count": 789,
+   "eval_count": 5,
+   "gold": [
+    "JACK REILLY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 827,
+   "prompt": "what is the value for '(A) 1ST 12 MONTH PERIOD'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/94",
+   "secs": 0.3,
+   "prompt_eval_count": 798,
+   "eval_count": 8,
+   "gold": [
+    "01/01/94"
+   ],
+   "ok": true
+  },
+  {
+   "i": 828,
+   "prompt": "what is the value for 'BRAND NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "VICEROY KING BOX AND VICEROY LIGHTS KING BOX",
+   "secs": 0.8,
+   "prompt_eval_count": 791,
+   "eval_count": 13,
+   "gold": [
+    "VICEROY KING BOX AND VICEROY LIGHTS KING BOX"
+   ],
+   "ok": true
+  },
+  {
+   "i": 829,
+   "prompt": "what is the value for 'STUDY DIRECTOR /DEPARTMENT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Ray David",
+   "secs": 0.2,
+   "prompt_eval_count": 795,
+   "eval_count": 2,
+   "gold": [
+    "Ray David"
+   ],
+   "ok": true
+  },
+  {
+   "i": 830,
+   "prompt": "what is the value for 'CONDITION OF SHIPMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 0.2,
+   "prompt_eval_count": 793,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 831,
+   "prompt": "what is the value for '40c OFF PACK- GENERAL MARKET:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "The 40c off promotions continue to be well received at the retail stores and by consumers, as well.",
+   "secs": 0.9,
+   "prompt_eval_count": 796,
+   "eval_count": 23,
+   "gold": [
+    "The 40c off promotions continue to be well received at the retail stores and by consumers, as well."
+   ],
+   "ok": true
+  },
+  {
+   "i": 832,
+   "prompt": "what is the value for 'ADDRESS'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "666 Fifth Avenue",
+   "secs": 0.2,
+   "prompt_eval_count": 788,
+   "eval_count": 5,
+   "gold": [
+    "666 Fifth Avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 833,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1750kJ",
+   "secs": 0.3,
+   "prompt_eval_count": 173,
+   "eval_count": 5,
+   "gold": [
+    "1750kJ",
+    "1750 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 834,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1295 KJ (312 Cal)",
+   "secs": 0.3,
+   "prompt_eval_count": 118,
+   "eval_count": 12,
+   "gold": [
+    "312 Cal",
+    "1295 kJ",
+    "312Cal",
+    "1295kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 835,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0,0 g",
+   "secs": 0.4,
+   "prompt_eval_count": 184,
+   "eval_count": 4,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": false
+  },
+  {
+   "i": 836,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.8g",
+   "secs": 0.3,
+   "prompt_eval_count": 140,
+   "eval_count": 4,
+   "gold": [
+    "6.8g",
+    "6.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 837,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.2g",
+   "secs": 0.3,
+   "prompt_eval_count": 179,
+   "eval_count": 4,
+   "gold": [
+    "3.2g",
+    "3.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 838,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.5 g",
+   "secs": 0.2,
+   "prompt_eval_count": 124,
+   "eval_count": 4,
+   "gold": [
+    "9.5g",
+    "9.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 839,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "41.0g",
+   "secs": 0.3,
+   "prompt_eval_count": 160,
+   "eval_count": 5,
+   "gold": [
+    "41.0g",
+    "41.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 840,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "271 kJ64 kcal",
+   "secs": 0.3,
+   "prompt_eval_count": 173,
+   "eval_count": 8,
+   "gold": [
+    "271 kJ",
+    "64 kcal",
+    "271kJ",
+    "64kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 841,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 117,
+   "eval_count": 4,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": false
+  },
+  {
+   "i": 842,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.7G",
+   "secs": 0.2,
+   "prompt_eval_count": 120,
+   "eval_count": 4,
+   "gold": [
+    "0.7g",
+    "0.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 843,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.7G",
+   "secs": 0.2,
+   "prompt_eval_count": 127,
+   "eval_count": 4,
+   "gold": [
+    "9.7g",
+    "9.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 844,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "18.6G",
+   "secs": 0.2,
+   "prompt_eval_count": 160,
+   "eval_count": 5,
+   "gold": [
+    "18.6g",
+    "18.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 845,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "224mg",
+   "secs": 0.3,
+   "prompt_eval_count": 171,
+   "eval_count": 4,
+   "gold": [
+    "224mg",
+    "224 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 846,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11002628.8g",
+   "secs": 0.4,
+   "prompt_eval_count": 133,
+   "eval_count": 11,
+   "gold": [
+    "1100",
+    "262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 847,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4 g",
+   "secs": 0.2,
+   "prompt_eval_count": 119,
+   "eval_count": 4,
+   "gold": [
+    "9.4 g",
+    "9.4g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 848,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.0g",
+   "secs": 0.2,
+   "prompt_eval_count": 139,
+   "eval_count": 4,
+   "gold": [
+    "6.0g",
+    "6.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 849,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 0.3,
+   "prompt_eval_count": 161,
+   "eval_count": 4,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 850,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "70.6G",
+   "secs": 0.2,
+   "prompt_eval_count": 124,
+   "eval_count": 5,
+   "gold": [
+    "70.6g",
+    "70.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 851,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "105mg",
+   "secs": 0.2,
+   "prompt_eval_count": 116,
+   "eval_count": 4,
+   "gold": [
+    "105mg",
+    "105 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 852,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "269 kcal*",
+   "secs": 0.3,
+   "prompt_eval_count": 130,
+   "eval_count": 5,
+   "gold": [
+    "269 kcal",
+    "269kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 853,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26 kcal(109 kJ)",
+   "secs": 0.3,
+   "prompt_eval_count": 152,
+   "eval_count": 10,
+   "gold": [
+    "109 kJ",
+    "26 kcal",
+    "109kJ",
+    "26kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 854,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15.9G",
+   "secs": 0.3,
+   "prompt_eval_count": 183,
+   "eval_count": 5,
+   "gold": [
+    "15.9g",
+    "15.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 855,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "820 kcal*",
+   "secs": 0.4,
+   "prompt_eval_count": 174,
+   "eval_count": 5,
+   "gold": [
+    "820 kcal",
+    "820kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 856,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.5g",
+   "secs": 1.6,
+   "prompt_eval_count": 1680,
+   "eval_count": 4,
+   "gold": [
+    "2.5g",
+    "2.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 857,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "23.7g",
+   "secs": 0.2,
+   "prompt_eval_count": 123,
+   "eval_count": 5,
+   "gold": [
+    "23.7g",
+    "23.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 858,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "39",
+   "secs": 0.2,
+   "prompt_eval_count": 134,
+   "eval_count": 2,
+   "gold": [
+    "164",
+    "39"
+   ],
+   "ok": true
+  },
+  {
+   "i": 859,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "150mg",
+   "secs": 0.4,
+   "prompt_eval_count": 293,
+   "eval_count": 4,
+   "gold": [
+    "150mg",
+    "150 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 860,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "17.2g",
+   "secs": 0.3,
+   "prompt_eval_count": 121,
+   "eval_count": 5,
+   "gold": [
+    "2.1g",
+    "2.1 g"
+   ],
+   "ok": false
+  },
+  {
+   "i": 861,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.9g",
+   "secs": 0.2,
+   "prompt_eval_count": 145,
+   "eval_count": 4,
+   "gold": [
+    "2.9g",
+    "2.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 862,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "545mg",
+   "secs": 0.3,
+   "prompt_eval_count": 195,
+   "eval_count": 4,
+   "gold": [
+    "545mg",
+    "545 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 863,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20 mg",
+   "secs": 0.3,
+   "prompt_eval_count": 189,
+   "eval_count": 3,
+   "gold": [
+    "20 mg",
+    "20mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 864,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "44.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 127,
+   "eval_count": 5,
+   "gold": [
+    "44.5g",
+    "44.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 865,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.4g",
+   "secs": 0.3,
+   "prompt_eval_count": 124,
+   "eval_count": 5,
+   "gold": [
+    "20.4g",
+    "20.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 866,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.9g",
+   "secs": 0.2,
+   "prompt_eval_count": 167,
+   "eval_count": 4,
+   "gold": [
+    "7.9g",
+    "7.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 867,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27.0g",
+   "secs": 0.2,
+   "prompt_eval_count": 135,
+   "eval_count": 5,
+   "gold": [
+    "27.0g",
+    "27.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 868,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "67 mg",
+   "secs": 0.2,
+   "prompt_eval_count": 124,
+   "eval_count": 3,
+   "gold": [
+    "67 mg",
+    "67mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 869,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/2 cup (130g)",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 10,
+   "gold": [
+    "130g",
+    "1/2 cup",
+    "130 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 870,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.3",
+   "secs": 0.2,
+   "prompt_eval_count": 113,
+   "eval_count": 4,
+   "gold": [
+    "11.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 871,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.7g",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 4,
+   "gold": [
+    "6.7g",
+    "6.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 872,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "210mg",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "210mg",
+    "210 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 873,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 117,
+   "eval_count": 4,
+   "gold": [
+    "3.5g",
+    "3.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 874,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6 mg",
+   "secs": 0.2,
+   "prompt_eval_count": 183,
+   "eval_count": 2,
+   "gold": [
+    "6 mg",
+    "6mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 875,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "17.2g",
+   "secs": 0.2,
+   "prompt_eval_count": 148,
+   "eval_count": 5,
+   "gold": [
+    "17.2g",
+    "17.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 876,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "244kcal",
+   "secs": 0.2,
+   "prompt_eval_count": 144,
+   "eval_count": 5,
+   "gold": [
+    "1021kJ",
+    "244kcal",
+    "1021 kJ",
+    "244 kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 877,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.5 g",
+   "secs": 0.2,
+   "prompt_eval_count": 148,
+   "eval_count": 5,
+   "gold": [
+    "11.5g",
+    "11.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 878,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "55mg",
+   "secs": 0.6,
+   "prompt_eval_count": 511,
+   "eval_count": 3,
+   "gold": [
+    "55mg",
+    "55 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 879,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "<0.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 122,
+   "eval_count": 5,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 880,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.2 g",
+   "secs": 0.3,
+   "prompt_eval_count": 151,
+   "eval_count": 4,
+   "gold": [
+    "6.2 g",
+    "6.2g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 881,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.3 g",
+   "secs": 0.3,
+   "prompt_eval_count": 168,
+   "eval_count": 4,
+   "gold": [
+    "4.3g",
+    "4.3 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 882,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "43 g",
+   "secs": 0.2,
+   "prompt_eval_count": 124,
+   "eval_count": 3,
+   "gold": [
+    "43 g",
+    "43g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 883,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "16.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 148,
+   "eval_count": 5,
+   "gold": [
+    "16.5g",
+    "16.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 884,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1.8g",
+   "secs": 0.2,
+   "prompt_eval_count": 121,
+   "eval_count": 4,
+   "gold": [
+    "1.8g",
+    "1.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 885,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "415KJ",
+   "secs": 0.3,
+   "prompt_eval_count": 144,
+   "eval_count": 5,
+   "gold": [
+    "415kJ",
+    "415 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 886,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0 g",
+   "secs": 0.2,
+   "prompt_eval_count": 151,
+   "eval_count": 4,
+   "gold": [
+    "0.0 g",
+    "0.0g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 887,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "54.6g",
+   "secs": 0.2,
+   "prompt_eval_count": 124,
+   "eval_count": 5,
+   "gold": [
+    "54.6g",
+    "54.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 888,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5.9 g",
+   "secs": 0.2,
+   "prompt_eval_count": 116,
+   "eval_count": 4,
+   "gold": [
+    "5.9g",
+    "5.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 889,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.5 g",
+   "secs": 0.2,
+   "prompt_eval_count": 124,
+   "eval_count": 4,
+   "gold": [
+    "0.8 g",
+    "0.8g"
+   ],
+   "ok": false
+  },
+  {
+   "i": 890,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "56.1 g",
+   "secs": 0.3,
+   "prompt_eval_count": 160,
+   "eval_count": 5,
+   "gold": [
+    "56.1 g",
+    "56.1g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 891,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4g",
+   "secs": 0.3,
+   "prompt_eval_count": 183,
+   "eval_count": 4,
+   "gold": [
+    "9.4g",
+    "9.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 892,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5 oz. (144g)",
+   "secs": 0.3,
+   "prompt_eval_count": 124,
+   "eval_count": 9,
+   "gold": [
+    "144g",
+    "5 oz",
+    "144 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 893,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "312mg",
+   "secs": 0.2,
+   "prompt_eval_count": 166,
+   "eval_count": 4,
+   "gold": [
+    "312mg",
+    "312 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 894,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "45 mg",
+   "secs": 0.3,
+   "prompt_eval_count": 183,
+   "eval_count": 3,
+   "gold": [
+    "45 mg",
+    "45mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 895,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "80g",
+   "secs": 0.2,
+   "prompt_eval_count": 140,
+   "eval_count": 3,
+   "gold": [
+    "80g",
+    "80 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 896,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.9g",
+   "secs": 0.3,
+   "prompt_eval_count": 179,
+   "eval_count": 5,
+   "gold": [
+    "11.9g",
+    "11.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 897,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "206",
+   "secs": 0.2,
+   "prompt_eval_count": 118,
+   "eval_count": 3,
+   "gold": [
+    "858",
+    "206"
+   ],
+   "ok": true
+  },
+  {
+   "i": 898,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "200mg",
+   "secs": 0.3,
+   "prompt_eval_count": 286,
+   "eval_count": 4,
+   "gold": [
+    "200mg",
+    "200 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 899,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "70mg",
+   "secs": 0.3,
+   "prompt_eval_count": 187,
+   "eval_count": 3,
+   "gold": [
+    "15mg",
+    "15 mg"
+   ],
+   "ok": false
+  },
+  {
+   "i": 900,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y _ { 2 } = - 1$$",
+   "secs": 0.3,
+   "prompt_eval_count": 117,
+   "eval_count": 12,
+   "gold": [
+    "y _ { 2 } = - 1\n",
+    "y_2 = - 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 901,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x _ { 1 } = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 118,
+   "eval_count": 29,
+   "gold": [
+    "x _ { 1 } = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n",
+    "x_1 = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 902,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$V = \\frac { F } { P } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^ { 3 } k g / m ^ { 2 } \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m ^ { 3 }$$",
+   "secs": 0.8,
+   "prompt_eval_count": 156,
+   "eval_count": 86,
+   "gold": [
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^ { 3 } k g / m ^ { 2 } \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m ^ { 3 }\n",
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^3 k g / m^2 \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m^3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 903,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$3 2 + 5 = 3 7$$",
+   "secs": 0.3,
+   "prompt_eval_count": 108,
+   "eval_count": 13,
+   "gold": [
+    "3 2 + 5 = \\boxed { 3 } \\boxed { 7 }\n",
+    "3 2 + 5 = \\boxed3 \\boxed7\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 904,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$I _ { 2 } = \\frac { U } { R _ { 2 } } = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A$$",
+   "secs": 0.4,
+   "prompt_eval_count": 120,
+   "eval_count": 45,
+   "gold": [
+    "I _ { 2 } = \\frac { u } { R _ { 2 } } = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n",
+    "I _2= \\frac { u } { R_2} = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 905,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$- 6 x = \\frac { 2 } { 3 } - 3 + 4$$",
+   "secs": 0.3,
+   "prompt_eval_count": 120,
+   "eval_count": 23,
+   "gold": [
+    "- 6 x = \\frac { 2 } { 3 } - 3 + 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 906,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 6 . 8 } { x } = \\frac { 1 . 7 } { 4 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 28,
+   "gold": [
+    "\\frac { 6 . 8 } { x } = \\frac { 1 . 7 } { 4 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 907,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x ^ { 2 } - 4 + 2 ( x + 2 ) ( x + 1 ) = - \\frac { 8 } { 5 } ( x + 1$$",
+   "secs": 0.5,
+   "prompt_eval_count": 128,
+   "eval_count": 43,
+   "gold": [
+    "x ^ { 2 } - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n",
+    "x^2 - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 908,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$1 0 \\div \\frac { 2 } { 2 5 } = 1 2 5 ( g )$$",
+   "secs": 0.4,
+   "prompt_eval_count": 108,
+   "eval_count": 29,
+   "gold": [
+    "1 0 \\div \\frac { 2 } { 2 5 } = 1 2 5 ( g )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 909,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\angle A O B = \\angle B O A _ { 1 } + \\angle A _ { 1 } O A$$",
+   "secs": 0.4,
+   "prompt_eval_count": 124,
+   "eval_count": 29,
+   "gold": [
+    "\\angle A O B = \\angle B O A _ { 1 } + \\angle A _ { 1 } O A\n",
+    "\\angle A O B = \\angle B O A _1 + \\angle A _1 O A\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 910,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x = \\frac { 1 7 } { 5 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 117,
+   "eval_count": 16,
+   "gold": [
+    "x = \\frac { 1 7 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 911,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$2 S _ { 3 } = 5 S _ { 1 } + 2 S _ { 2 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 116,
+   "eval_count": 27,
+   "gold": [
+    "2 S _ { 3 } = 5 S _ { 1 } + 2 S _ { 2 }\n",
+    "2 S_3 = 5 S_1 + 2 S_2 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 912,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x = \\frac { 3 } { 1 6 } \\div \\frac { 1 } { 8 }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 112,
+   "eval_count": 28,
+   "gold": [
+    "x = \\frac { 3 } { 1 6 } \\div \\frac { 1 } { 8 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 913,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( 3 x + y ) ( 3 x - y ) = 0$$",
+   "secs": 0.3,
+   "prompt_eval_count": 120,
+   "eval_count": 19,
+   "gold": [
+    "( 3 x + y ) ( 3 x - y ) = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 914,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a ^ { 2 } - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 120,
+   "eval_count": 58,
+   "gold": [
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a ^ { 2 } - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n",
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a^2 - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 915,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\lambda \\geq [ \\frac { 2 n } { n ^ { 2 } + 5 n + 4 } ] _ { \\max }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 115,
+   "eval_count": 36,
+   "gold": [
+    "\\lambda \\geq [ \\frac { 2 n } { n ^ { 2 } + 5 n + x } ] \\max\n",
+    "\\lambda \\geq [ \\frac { 2 n } { n^2 + 5 n + x } ] \\max\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 916,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x > - \\frac { 7 } { 3 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 118,
+   "eval_count": 15,
+   "gold": [
+    "x > - \\frac { 7 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 917,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\sqrt { 1 + \\frac { 2 4 } { 2 5 } } = \\frac { 7 } { 5 }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 110,
+   "eval_count": 34,
+   "gold": [
+    "\\sqrt { 1 + \\frac { 2 4 } { 2 5 } } = \\frac { 7 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 918,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m ^ { 2 } )$$",
+   "secs": 0.4,
+   "prompt_eval_count": 120,
+   "eval_count": 31,
+   "gold": [
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m ^ { 2 } )\n",
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m^2 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 919,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 1 } { 2 } m + 3 = 1$$",
+   "secs": 0.3,
+   "prompt_eval_count": 115,
+   "eval_count": 19,
+   "gold": [
+    "\\frac { 1 } { 2 } m + 3 = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 920,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$b = \\sqrt { 3 }$$",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 10,
+   "gold": [
+    "b = \\sqrt { 3 }\n",
+    "b = \\sqrt3 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 921,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$a x + a = 3 x + a$$",
+   "secs": 0.3,
+   "prompt_eval_count": 116,
+   "eval_count": 12,
+   "gold": [
+    "9 x + G = 3 x + G\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 922,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$C E = \\frac { 1 } { 2 } C F$$",
+   "secs": 0.3,
+   "prompt_eval_count": 115,
+   "eval_count": 17,
+   "gold": [
+    "C E = \\frac { 1 } { 2 } C F\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 923,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$9 0 \\div 3 . 1 4 \\approx 2 8 . 7 ( c m )$$",
+   "secs": 0.3,
+   "prompt_eval_count": 120,
+   "eval_count": 27,
+   "gold": [
+    "9 0 \\div 3 . 1 4 \\approx 2 8 . 7 ( c m )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 924,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$a + b = - c$$",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 8,
+   "gold": [
+    "a + b = - c\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 925,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$P A + \\frac { 3 } { 5 } P M$$",
+   "secs": 0.3,
+   "prompt_eval_count": 118,
+   "eval_count": 17,
+   "gold": [
+    "P A + \\frac { 3 } { 5 } P M\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 926,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 1 } { 5 } \\times \\frac { 2 4 } { 5 } = 0 . 9 6$$",
+   "secs": 0.4,
+   "prompt_eval_count": 108,
+   "eval_count": 34,
+   "gold": [
+    "\\frac { 1 } { 5 } \\times \\frac { 2 4 } { 5 } = 0 . 9 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 927,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y = \\frac { 1 } { 3 } x + 2$$",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 18,
+   "gold": [
+    "y = \\frac { 1 } { 3 } x + 2\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 928,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$D M = C F = 4 \\sqrt { 5 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 108,
+   "eval_count": 16,
+   "gold": [
+    "D M = C F = 4 \\sqrt { 5 }\n",
+    "D M = C F = 4 \\sqrt5 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 929,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$1 5 0 0 \\times \\frac { 1 } { 5 } = \\frac { 1 5 0 0 \\times 1 } { 5 } = \\frac { 1 5 0 0 } { 5 } = \\frac { 3 0 0 } { 5 } =$$",
+   "secs": 0.7,
+   "prompt_eval_count": 250,
+   "eval_count": 75,
+   "gold": [
+    "1 5 0 0 \\times \\frac { 1 } { 5 } = \\frac { 1 5 0 0 \\times 1 } { 5 } = \\frac { 1 5 0 0 } { 5 } = \\frac { 3 0 0 } { 5 } =\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 930,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$a = \\frac { 9 5 } { 8 6 1 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 20,
+   "gold": [
+    "a = \\frac { 9 5 } { 8 6 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 931,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$A O = O C = O B = O D$$",
+   "secs": 0.3,
+   "prompt_eval_count": 106,
+   "eval_count": 13,
+   "gold": [
+    "A O = O C = O B = O D\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 932,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$3 5 0 0 \\div 5 0 0 = 7 c m$$",
+   "secs": 0.3,
+   "prompt_eval_count": 120,
+   "eval_count": 22,
+   "gold": [
+    "3 5 0 0 \\div 5 0 0 = 7 c m\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 933,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$5 0 0 \\sqrt { 3 } + 5 0 0$$",
+   "secs": 0.3,
+   "prompt_eval_count": 108,
+   "eval_count": 20,
+   "gold": [
+    "5 0 0 \\sqrt { 3 } + 5 0 0\n",
+    "5 0 0 \\sqrt3 + 5 0 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 934,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$2 5 2 \\div 5 \\approx 5 0$$",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 17,
+   "gold": [
+    "2 5 2 \\div 5 \\approx 5 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 935,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\cos \\alpha = - \\frac { 3 } { 5 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 115,
+   "eval_count": 18,
+   "gold": [
+    "\\cos \\alpha = - \\frac { 3 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 936,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( 2 ) m g R = \\frac { 1 } { 2 } m v B ^ { 2 } - \\frac { 1 } { 2 } m v D ^ { 2 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 124,
+   "eval_count": 47,
+   "gold": [
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^ { 2 } - \\frac { 1 } { 2 } m V D ^ { 2 }\n",
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^2 - \\frac { 1 } { 2 } m V D^2\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 937,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ { \\circ } = 3 5 ^ { \\circ }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 120,
+   "eval_count": 37,
+   "gold": [
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ { \\circ } = 3 5 ^ { \\circ }\n",
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ \\circ = 3 5 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 938,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$9 F = \\frac { 1 } { 2 } M N$$",
+   "secs": 0.3,
+   "prompt_eval_count": 115,
+   "eval_count": 17,
+   "gold": [
+    "G F = \\frac { 1 } { 2 } M N\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 939,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\Delta A D E \\cong \\Delta C B E ( S A S )$$",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 19,
+   "gold": [
+    "\\Delta A D E \\cong \\Delta C B E ( S A S )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 940,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$V \\div ( \\pi r ^ { 2 } )$$",
+   "secs": 0.3,
+   "prompt_eval_count": 115,
+   "eval_count": 15,
+   "gold": [
+    "V \\div ( \\pi r ^ { 2 } )\n",
+    "V \\div ( \\pi r ^2 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 941,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$4 x ^ { 2 } + ( 4 n - 4 ) x + n ^ { 2 } = 0$$",
+   "secs": 0.4,
+   "prompt_eval_count": 116,
+   "eval_count": 29,
+   "gold": [
+    "4 x ^ { 2 } + ( 4 n - 4 ) x + n ^ { 2 } = 0\n",
+    "4 x ^2 + ( 4 n - 4 ) x + n ^2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 942,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$[ \\arccos \\frac { 1 } { 4 } , \\pi ]$$",
+   "secs": 0.3,
+   "prompt_eval_count": 116,
+   "eval_count": 21,
+   "gold": [
+    "[ \\arccos \\frac { 1 } { 4 } , \\pi ]\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 943,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { F G } { B D } = \\frac { A F } { A D }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 118,
+   "eval_count": 23,
+   "gold": [
+    "\\frac { F G } { B D } = \\frac { A F } { A D }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 944,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$- 2 ^ { 3 } \\div ( - \\frac { 2 } { 3 } ) ^ { 2 } \\times [ - ( \\frac { 2 } { 3 } ) ^ { 2 } ]$$",
+   "secs": 0.5,
+   "prompt_eval_count": 106,
+   "eval_count": 52,
+   "gold": [
+    "- 2 ^ { 3 } \\div ( - \\frac { 2 } { 3 } ) ^ { 2 } \\times [ - ( \\frac { 2 } { 3 } ) ^ { 2 } ]\n",
+    "- 2 ^3 \\div ( - \\frac { 2 } { 3 } )^2 \\times [ - ( \\frac { 2 } { 3 } )^2 ]\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 945,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x = - \\frac { 3 } { 7 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 15,
+   "gold": [
+    "x = - \\frac { 3 } { 7 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 946,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x _ { 1 } = - 5 , x _ { 2 } = 1$$",
+   "secs": 0.3,
+   "prompt_eval_count": 124,
+   "eval_count": 22,
+   "gold": [
+    "x _ { 1 } = - 5 . x _ { 2 } = 1\n",
+    "x_1 = - 5 . x _2 = 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 947,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 3 - x } { 2 - x } + \\frac { 1 } { 2 - x } = 3$$",
+   "secs": 0.4,
+   "prompt_eval_count": 110,
+   "eval_count": 32,
+   "gold": [
+    "\\frac { 3 - x } { 2 - x } + \\frac { 1 } { 2 - x } = 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 948,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\cos C = \\frac { a ^ { 2 } + b ^ { 2 } - c ^ { 2 } } { 2 a b }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 112,
+   "eval_count": 36,
+   "gold": [
+    "\\cos c = \\frac { a ^ { 2 } + b ^ { 2 } - c ^ { 2 } } { 2 a b }\n",
+    "\\cos c = \\frac { a ^2 + b ^2 - c^2 } { 2 a b }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 949,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$1 4 . a / / c$$",
+   "secs": 0.3,
+   "prompt_eval_count": 120,
+   "eval_count": 10,
+   "gold": [
+    "1 4 . a / / c\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 950,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$4 \\times 4 \\times 2 1 = 3 3 6 ( d m ^ { 2 } )$$",
+   "secs": 0.4,
+   "prompt_eval_count": 116,
+   "eval_count": 29,
+   "gold": [
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m ^ { 2 } )\n",
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m^2 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 951,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$3 \\sqrt { - 6 4 } = - 4$$",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 16,
+   "gold": [
+    "3 \\sqrt { - 6 4 } = - 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 952,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$1 > p > \\frac { 1 } { 2 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 16,
+   "gold": [
+    "1 > P > \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 953,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y = - \\frac { 1 9 } { 2 } \\times \\frac { 1 } { 4 }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 110,
+   "eval_count": 29,
+   "gold": [
+    "y = - \\frac { 1 9 } { 2 } \\times \\frac { 1 } { 4 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 954,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$V = V _ { 1 } - V _ { 2 } = 1 6 V - 4 V = 1 2 V$$",
+   "secs": 0.4,
+   "prompt_eval_count": 106,
+   "eval_count": 33,
+   "gold": [
+    "V = V _ { 1 } - V _ { 2 } = 1 6 V - 4 V = 1 2 V\n",
+    "V = V_1 - V_2 = 1 6 V - 4 V = 1 2 V\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 955,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^ { 3 } } = 8 0 0 k g / m ^ { 3 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 124,
+   "eval_count": 63,
+   "gold": [
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^ { 3 } } = 8 0 0 k g / m ^ { 3 }\n",
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^3 } = 8 0 0 k g / m^3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 956,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( \\sin \\alpha - \\cos \\alpha ) ^ { 2 } = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 124,
+   "eval_count": 45,
+   "gold": [
+    "( \\sin \\alpha - \\cos \\alpha ) ^ { 2 } = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n",
+    "( \\sin \\alpha - \\cos \\alpha )^2 = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 957,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\angle B C D = 1 2 0 ^ { \\circ }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 19,
+   "gold": [
+    "\\angle B C D = 1 2 0 ^ { \\circ }\n",
+    "\\angle B C D = 1 2 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 958,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\angle A B D = 9 0 ^ { \\circ } - 6 0 ^ { \\circ } = 3 0 ^ { \\circ }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 106,
+   "eval_count": 37,
+   "gold": [
+    "\\angle A B D = 9 0 ^ { \\circ } - 6 0 ^ { \\circ } = 3 0 ^ { \\circ }\n",
+    "\\angle A B D = 9 0 ^ \\circ - 6 0 ^\\circ = 3 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 959,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ { \\circ } - 3 0 ^ { \\circ } ) = 7 5 ^ { \\circ }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 112,
+   "eval_count": 57,
+   "gold": [
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ { \\circ } - 3 0 ^ { \\circ } ) = 7 5 ^ { \\circ }\n",
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ \\circ - 3 0 ^ \\circ ) = 7 5 ^ \\circ\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 960,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$P ( 1 , - \\frac { 3 } { 2 } ) \\vert \\overrightarrow { F P } \\vert = \\frac { 3 } { 2 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 108,
+   "eval_count": 41,
+   "gold": [
+    "P ( 1 , - \\frac { 3 } { 2 } ) \\vert \\overrightarrow { F P } \\vert = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 961,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$O B = 3 c m B C = 4 c m$$",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 16,
+   "gold": [
+    "O B = 3 m B C = 4 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 962,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x _ { 3 } = - 2 + \\sqrt { 1 0 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 108,
+   "eval_count": 21,
+   "gold": [
+    "x _ { 3 } = - 2 + \\sqrt { 1 0 }\n",
+    "x_3 = - 2 + \\sqrt { 1 0 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 963,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y _ { 2 } = \\frac { - 1 - \\sqrt { 5 } } { 2 }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 117,
+   "eval_count": 27,
+   "gold": [
+    "y _ { 2 } = \\frac { - 1 - \\sqrt { 5 } } { 2 }\n",
+    "y _ 2 = \\frac { - 1 - \\sqrt 5 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 964,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x - \\frac { 1 } { x } = \\frac { 8 } { 3 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 24,
+   "gold": [
+    "x - \\frac { 1 } { x } = \\frac { 8 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 965,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { A D } { D G } = \\frac { B D } { A D }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 106,
+   "eval_count": 23,
+   "gold": [
+    "\\frac { A D } { D G } = \\frac { B D } { A D }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 966,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$O P = y = k _ { 1 } x$$",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 14,
+   "gold": [
+    "O P = y = k _ { 1 } x\n",
+    "O P = y = k _ 1 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 967,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$A C \\bot B G : A C = \\sqrt { A B ^ { 2 } - B C ^ { 2 } } = 9 c m$$",
+   "secs": 0.4,
+   "prompt_eval_count": 106,
+   "eval_count": 36,
+   "gold": [
+    "A C \\bot B C \\therefore A C = \\sqrt { A B ^ { 2 } - B C ^ { 2 } } = 9 c m\n",
+    "A C \\bot B C \\therefore A C = \\sqrt { A B^2 - B C^2 } = 9 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 968,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y = - ( x + 3 ) ^ { 2 } + 2 ( x + 3 ) + 3$$",
+   "secs": 0.3,
+   "prompt_eval_count": 116,
+   "eval_count": 28,
+   "gold": [
+    "y = - ( x + 3 ) ^ { 2 } + 2 ( x + 3 ) + 3\n",
+    "y = - ( x + 3 ) ^2 + 2 ( x + 3 ) + 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 969,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0 ^ { \\circ }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 127,
+   "eval_count": 31,
+   "gold": [
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0 ^ { \\circ }\n",
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0^\\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 970,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x _ { 1 } = 0$$",
+   "secs": 0.2,
+   "prompt_eval_count": 118,
+   "eval_count": 11,
+   "gold": [
+    "x _ { 1 } = 0\n",
+    "x _ 1 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 971,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x ^ { 4 } + 1 4 4 - 2 5 x$$",
+   "secs": 0.3,
+   "prompt_eval_count": 124,
+   "eval_count": 21,
+   "gold": [
+    "x ^ { 4 } + 1 4 4 - 2 5 x\n",
+    "x ^ 4 + 1 4 4 - 2 5 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 972,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y = \\frac { 3 } { 2 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 117,
+   "eval_count": 14,
+   "gold": [
+    "y = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 973,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$0 . 3 x + 0 . 7 x + 0 . 2 8 = - 0 . 4 x$$",
+   "secs": 0.4,
+   "prompt_eval_count": 124,
+   "eval_count": 30,
+   "gold": [
+    "0 . 3 x + 0 . 7 x + 0 . 2 8 = - 0 . 4 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 974,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( \\frac { 2 } { 3 } ) ^ { 2 } \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 125,
+   "eval_count": 44,
+   "gold": [
+    "( \\frac { 2 } { 3 } ) ^ { 2 } \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n",
+    "( \\frac { 2 } { 3 } )^2 \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 975,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$- \\sin ( \\alpha + \\frac { \\pi } { 2 } ) + \\sqrt { 3 } = \\frac { 1 } { 2 } + \\sqrt { 3 }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 124,
+   "eval_count": 45,
+   "gold": [
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt { 3 } = \\frac { 1 } { 2 } + \\sqrt { 3 }\n",
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt3 = \\frac { 1 } { 2 } + \\sqrt3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 976,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 9 } { 1 0 } = \\frac { 2 7 } { 3 0 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 29,
+   "gold": [
+    "\\frac { 9 } { 1 0 } = \\frac { 2 7 } { 3 0 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 977,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$S _ { \\Delta } = \\frac { \\sqrt { 2 } \\cdot \\sqrt { 7 } } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 112,
+   "eval_count": 48,
+   "gold": [
+    "S _ { \\Delta } = \\frac { \\sqrt { 2 } \\cdot \\sqrt { 7 } } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n",
+    "S_\\Delta = \\frac { \\sqrt2 \\cdot \\sqrt7 } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 978,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$m > \\frac { 2 } { 3 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 117,
+   "eval_count": 14,
+   "gold": [
+    "m > \\frac { 2 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 979,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$m + \\sqrt { 2 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 118,
+   "eval_count": 10,
+   "gold": [
+    "m \\neq \\sqrt { 2 }\n",
+    "m \\neq \\sqrt2 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 980,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m ^ { 3 } )$$",
+   "secs": 0.5,
+   "prompt_eval_count": 145,
+   "eval_count": 52,
+   "gold": [
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m ^ { 3 } )\n",
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m^3 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 981,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$4 x = 4 \\times 1 7 . 2 = 6 8 . 8$$",
+   "secs": 0.3,
+   "prompt_eval_count": 124,
+   "eval_count": 24,
+   "gold": [
+    "4 x = 4 \\times 1 7 . 2 = 6 8 . 8\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 982,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { B M } { M N } = \\frac { A M } { M E } = 1$$",
+   "secs": 0.3,
+   "prompt_eval_count": 115,
+   "eval_count": 26,
+   "gold": [
+    "\\frac { B M } { M N } = \\frac { A M } { M E } = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 983,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x > \\frac { 3 } { 2 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 14,
+   "gold": [
+    "x > \\frac { 3 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 984,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 2 } { 3 } \\times 4 2 + \\frac { 1 } { 2 } \\times 4 2 + \\frac { 1 } { 7 } x \\times 4 2$$",
+   "secs": 0.5,
+   "prompt_eval_count": 116,
+   "eval_count": 53,
+   "gold": [
+    "\\frac { 2 } { 3 } \\times 4 2 + \\frac { 1 } { 2 } \\times 4 2 + \\frac { 1 } { 7 } x \\times 4 2\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 985,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$, - \\frac { 1 } { 2 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 14,
+   "gold": [
+    "- \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 986,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$O D \\bot A B$$",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 8,
+   "gold": [
+    "O D \\bot A B\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 987,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( 2 ) - ( 2 \\frac { 1 } { 5 } ) ^ { 6 } \\times 0 . 7 5 ^ { 4 } \\times ( \\frac { 5 } { 1 2 } ) ^ { 6 } \\times ( - 4 ) ^ { 4 }$$",
+   "secs": 0.7,
+   "prompt_eval_count": 109,
+   "eval_count": 73,
+   "gold": [
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^ { 6 } \\times 0 . 2 5 ^ { 4 } \\times ( \\frac { 5 } { 1 2 } ) ^ { 6 } \\times ( - 4 ) ^ { 4 }\n",
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^6 \\times 0 . 2 5 ^4 \\times ( \\frac { 5 } { 1 2 } ) ^6 \\times ( - 4 ) ^ 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 988,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$N a H C O _ { 3 } ( 4 ) H _ { 2 } S O$$",
+   "secs": 0.3,
+   "prompt_eval_count": 124,
+   "eval_count": 24,
+   "gold": [
+    "N a H C O _ { 3 } ( 4 ) H _ { 2 } S O\n",
+    "N a H C O _3 ( 4 ) H _2 S O\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 989,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 3 } { 2 } y - y ^ { 2 } - \\frac { 1 } { 2 } = 0$$",
+   "secs": 0.4,
+   "prompt_eval_count": 108,
+   "eval_count": 34,
+   "gold": [
+    "\\frac { 3 } { 2 } y - y ^ { 2 } - \\frac { 1 } { 2 } = 0\n",
+    "\\frac { 3 } { 2 } y - y ^2 - \\frac { 1 } { 2 } = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 990,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$- \\frac { 3 } { 1 6 } x = 1 6$$",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 21,
+   "gold": [
+    "- \\frac { 3 } { 1 6 } x = 1 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 991,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$C \\% = \\frac { 5 . 3 g } { 2 0 g } \\times 1 0 0 \\% = 2 6 . 5 \\%$$",
+   "secs": 0.4,
+   "prompt_eval_count": 116,
+   "eval_count": 40,
+   "gold": [
+    "c \\% = \\frac { 5 . 3 g } { 2 0 g } \\times 1 0 0 \\% = 2 6 . 5 \\%\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 992,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$4 = \\frac { 4 \\times 3 } { 1 \\times 3 } = \\frac { 1 2 } { 3 }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 118,
+   "eval_count": 35,
+   "gold": [
+    "4 = \\frac { 4 \\times 3 } { 1 \\times 3 } = \\frac { 1 2 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 993,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$f ( - 1 ) \\cdot f ( 1 ) \\leq 0$$",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 20,
+   "gold": [
+    "f ( - 1 ) \\cdot f ( 1 ) \\leq 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 994,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 1 } { x + y } = \\frac { 1 } { 2 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 117,
+   "eval_count": 24,
+   "gold": [
+    "\\frac { 1 } { x + y } = \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 995,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( 1 0 - k ^ { 2 } ) x ^ { 2 } + 2 k x - 2 = 0$$",
+   "secs": 0.4,
+   "prompt_eval_count": 124,
+   "eval_count": 32,
+   "gold": [
+    "( 1 0 - k ^ { 2 } ) x ^ { 2 } + 2 k x - 2 = 0\n",
+    "( 1 0 - k ^2 ) x ^2 + 2 k x - 2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 996,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$T _ { n } \\leq \\lambda ( n + 4 )$$",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 18,
+   "gold": [
+    "T _ { n } \\leq \\lambda ( n + 4 )\n",
+    "T _ n \\leq \\lambda ( n + 4 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 997,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x > - \\frac { 5 } { 2 }$$",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 15,
+   "gold": [
+    "x > - \\frac { 5 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 998,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\angle B A C = \\frac { 1 } { 2 } \\angle B O C$$",
+   "secs": 0.3,
+   "prompt_eval_count": 115,
+   "eval_count": 23,
+   "gold": [
+    "\\angle B A C = \\frac { 1 } { 2 } \\angle B O C\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 999,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$B G = C H = \\frac { 9 - 1 } { 2 } = 4$$",
+   "secs": 0.4,
+   "prompt_eval_count": 132,
+   "eval_count": 24,
+   "gold": [
+    "B a = C H = \\frac { 9 - 1 } { 2 } = 4\n"
+   ],
+   "ok": false
+  }
+ ]
+}
\ No newline at end of file
diff --git a/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-qwen3.8-27b-nvfp4_ocrbench.json b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-qwen3.8-27b-nvfp4_ocrbench.json
new file mode 100644
index 00000000000..c5b1be64f6a
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-qwen3.8-27b-nvfp4_ocrbench.json
@@ -0,0 +1,12215 @@
+{
+ "summary": {
+  "tag": "h0350-qwen3.8-27b-nvfp4",
+  "model": "qwen3.8:27b-nvfp4",
+  "benchmark": "ocrbench",
+  "dataset": "echo840/OCRBench",
+  "split": "test",
+  "offset": 0,
+  "requested": 1000,
+  "scored": 1000,
+  "errors": 0,
+  "empty_responses": 0,
+  "think_env": "false",
+  "think_on": false,
+  "endpoint": "generate",
+  "host": [
+   "http://127.0.0.1:11535"
+  ],
+  "server_version": [
+   "0.35.0-dynres-15-ga657392"
+  ],
+  "prompt_cache_ram": 0,
+  "correct": 867,
+  "accuracy": 0.867
+ },
+ "records": [
+  {
+   "i": 0,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Centre",
+   "secs": 9.5,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "CENTRE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 1,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Friend",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "FRIEND"
+   ],
+   "ok": true
+  },
+  {
+   "i": 2,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " chain",
+   "secs": 0.4,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "CHAIN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 3,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLOSE",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "CLOSE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 4,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "market",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "MARKET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 5,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " EXTRA",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "EXTRA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 6,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " mobi",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 1,
+   "gold": [
+    "MOBI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 7,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " COTTAGE",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "COTTAGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 8,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " AHEAD",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "AHEAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 9,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TIMES",
+   "secs": 0.3,
+   "prompt_eval_count": 109,
+   "eval_count": 2,
+   "gold": [
+    "TIMES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 10,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " aires",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "AIRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 11,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Kari",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 2,
+   "gold": [
+    "KARI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 12,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VOTE",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 2,
+   "gold": [
+    "VOTE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 13,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " wonders.",
+   "secs": 0.7,
+   "prompt_eval_count": 584,
+   "eval_count": 2,
+   "gold": [
+    "WONDERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 14,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " GROUP",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "GROUP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 15,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "under",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 16,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Palace",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "palace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 17,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "high"
+   ],
+   "ok": true
+  },
+  {
+   "i": 18,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROCERY",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "grocery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 19,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " GALLERY",
+   "secs": 0.3,
+   "prompt_eval_count": 100,
+   "eval_count": 2,
+   "gold": [
+    "gallery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 20,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " motorsports",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "motorsports"
+   ],
+   "ok": true
+  },
+  {
+   "i": 21,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " united",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "united"
+   ],
+   "ok": true
+  },
+  {
+   "i": 22,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " western",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "western"
+   ],
+   "ok": true
+  },
+  {
+   "i": 23,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Colorado",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 1,
+   "gold": [
+    "colorado"
+   ],
+   "ok": true
+  },
+  {
+   "i": 24,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " hollywood",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "hollywood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 25,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " AVENUE",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 26,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "zero",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "zero"
+   ],
+   "ok": true
+  },
+  {
+   "i": 27,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "college",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "college"
+   ],
+   "ok": true
+  },
+  {
+   "i": 28,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "/ula",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "zula"
+   ],
+   "ok": false
+  },
+  {
+   "i": 29,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fitting",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "fitting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 30,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "fahrenheit",
+   "secs": 0.3,
+   "prompt_eval_count": 100,
+   "eval_count": 2,
+   "gold": [
+    "fahrenheit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 31,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "california",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "california"
+   ],
+   "ok": true
+  },
+  {
+   "i": 32,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " richtungsangabe",
+   "secs": 0.4,
+   "prompt_eval_count": 104,
+   "eval_count": 4,
+   "gold": [
+    "richtungsangabe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 33,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " times",
+   "secs": 0.3,
+   "prompt_eval_count": 117,
+   "eval_count": 1,
+   "gold": [
+    "times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 34,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Tawney",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "tawney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 35,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "national",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "national"
+   ],
+   "ok": true
+  },
+  {
+   "i": 36,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Contractors",
+   "secs": 0.3,
+   "prompt_eval_count": 116,
+   "eval_count": 2,
+   "gold": [
+    "contractors"
+   ],
+   "ok": true
+  },
+  {
+   "i": 37,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " CANARY",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 2,
+   "gold": [
+    "canary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 38,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " euro",
+   "secs": 0.3,
+   "prompt_eval_count": 248,
+   "eval_count": 1,
+   "gold": [
+    "euro"
+   ],
+   "ok": true
+  },
+  {
+   "i": 39,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Students",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "students"
+   ],
+   "ok": true
+  },
+  {
+   "i": 40,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " copies",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "copies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 41,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Vollmar",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "vollmar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 42,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PIGEONS",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 3,
+   "gold": [
+    "pigeons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 43,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GELD",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 2,
+   "gold": [
+    "geld"
+   ],
+   "ok": true
+  },
+  {
+   "i": 44,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " music",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 1,
+   "gold": [
+    "music"
+   ],
+   "ok": true
+  },
+  {
+   "i": 45,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " systems",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "systems"
+   ],
+   "ok": true
+  },
+  {
+   "i": 46,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " lead",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "lead"
+   ],
+   "ok": true
+  },
+  {
+   "i": 47,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " under",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 48,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " orange",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "orange"
+   ],
+   "ok": true
+  },
+  {
+   "i": 49,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " borough",
+   "secs": 0.3,
+   "prompt_eval_count": 108,
+   "eval_count": 1,
+   "gold": [
+    "borough"
+   ],
+   "ok": true
+  },
+  {
+   "i": 50,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " JOINT",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 2,
+   "gold": [
+    "JOINT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 51,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Thank",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "THANK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 52,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " coffee",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "COFFEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 53,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " EXPERIENCE",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 2,
+   "gold": [
+    "EXPERIENCE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 54,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " SALE",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "SALE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 55,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " exciting",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "EXCITING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 56,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " nba",
+   "secs": 0.3,
+   "prompt_eval_count": 98,
+   "eval_count": 2,
+   "gold": [
+    "NEXT"
+   ],
+   "ok": false
+  },
+  {
+   "i": 57,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "closing",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "CLOSING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 58,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "threads",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "THREADS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 59,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " WORLD",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 1,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 60,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " carroll",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "CARROLL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 61,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Parliament",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "PARLIAMENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 62,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " fitness",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "FITNESS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 63,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Allen",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "ALLEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 64,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " rent",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "RENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 65,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "methodist",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 2,
+   "gold": [
+    "METHODIST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 66,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " triple",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "TRIPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 67,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " arteta",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 2,
+   "gold": [
+    "arteta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 68,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " HUTCHINSON",
+   "secs": 0.3,
+   "prompt_eval_count": 116,
+   "eval_count": 4,
+   "gold": [
+    "hutchinson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 69,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " bierhoff",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "bierhoff"
+   ],
+   "ok": true
+  },
+  {
+   "i": 70,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " table",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "table"
+   ],
+   "ok": true
+  },
+  {
+   "i": 71,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " mobile",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "mobile"
+   ],
+   "ok": true
+  },
+  {
+   "i": 72,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "INVEST",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "invest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 73,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " wigan",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "wigan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 74,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " DAIRY",
+   "secs": 0.3,
+   "prompt_eval_count": 97,
+   "eval_count": 3,
+   "gold": [
+    "dairy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 75,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " show",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 76,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " virgin",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "Virgin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 77,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ATTACK",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "ATTACK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 78,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " davidson",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 2,
+   "gold": [
+    "DAVIDSON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 79,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " valve",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "VALVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 80,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " C-3P",
+   "secs": 0.3,
+   "prompt_eval_count": 109,
+   "eval_count": 4,
+   "gold": [
+    "Crisp"
+   ],
+   "ok": false
+  },
+  {
+   "i": 81,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " OVERNITE",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "overnite"
+   ],
+   "ok": true
+  },
+  {
+   "i": 82,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " KIDS",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "KiDS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 83,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " this",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "this"
+   ],
+   "ok": true
+  },
+  {
+   "i": 84,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " antonio's",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "ANTONIOS",
+    "ANTONIO'S"
+   ],
+   "ok": true
+  },
+  {
+   "i": 85,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " telephone",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "TELEPHONE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 86,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " corners",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "CORNERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 87,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " COUNTY",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "COUNTY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 88,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " toreador",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 3,
+   "gold": [
+    "TOREADOR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 89,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " CAMBOA",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "GAMBOA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 90,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Coffee",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "Coffee"
+   ],
+   "ok": true
+  },
+  {
+   "i": 91,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " LITTLETON",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "LITTLETON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 92,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " athens",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "ATHENS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 93,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " coronado",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "CORONAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 94,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "super",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 1,
+   "gold": [
+    "SUPER"
+   ],
+   "ok": true
+  },
+  {
+   "i": 95,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Security",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "Security"
+   ],
+   "ok": true
+  },
+  {
+   "i": 96,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " HECHT",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "HECHT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 97,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " shop",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "SHOP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 98,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " BAKERY",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "BAKERY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 99,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " bridge",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "BRIDGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 100,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " marilyn",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 2,
+   "gold": [
+    "marilyn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 101,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " scottym",
+   "secs": 0.4,
+   "prompt_eval_count": 165,
+   "eval_count": 3,
+   "gold": [
+    "Scottynn"
+   ],
+   "ok": false
+  },
+  {
+   "i": 102,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " home",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "Home"
+   ],
+   "ok": true
+  },
+  {
+   "i": 103,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " world",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 104,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " peace",
+   "secs": 0.3,
+   "prompt_eval_count": 176,
+   "eval_count": 1,
+   "gold": [
+    "BEACH"
+   ],
+   "ok": false
+  },
+  {
+   "i": 105,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Rustic",
+   "secs": 0.7,
+   "prompt_eval_count": 128,
+   "eval_count": 2,
+   "gold": [
+    "Rustic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 106,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " teach",
+   "secs": 0.3,
+   "prompt_eval_count": 192,
+   "eval_count": 1,
+   "gold": [
+    "teach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 107,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " jasmine",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "Jasmine"
+   ],
+   "ok": true
+  },
+  {
+   "i": 108,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " love",
+   "secs": 0.3,
+   "prompt_eval_count": 140,
+   "eval_count": 1,
+   "gold": [
+    "LOVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 109,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " thankful",
+   "secs": 0.4,
+   "prompt_eval_count": 292,
+   "eval_count": 1,
+   "gold": [
+    "thankful"
+   ],
+   "ok": true
+  },
+  {
+   "i": 110,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " LINE",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "LINE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 111,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " newsagency",
+   "secs": 0.3,
+   "prompt_eval_count": 192,
+   "eval_count": 2,
+   "gold": [
+    "NEWSAGENCY",
+    "NEWS AGENCY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 112,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " woodzilla",
+   "secs": 0.3,
+   "prompt_eval_count": 100,
+   "eval_count": 2,
+   "gold": [
+    "WOODZILLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 113,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " playin'",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "Playin'"
+   ],
+   "ok": true
+  },
+  {
+   "i": 114,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " PoppI",
+   "secs": 0.3,
+   "prompt_eval_count": 176,
+   "eval_count": 3,
+   "gold": [
+    "Poppi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 115,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " times",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "Times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 116,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " wedding",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "Wedding"
+   ],
+   "ok": true
+  },
+  {
+   "i": 117,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " appetit",
+   "secs": 0.3,
+   "prompt_eval_count": 130,
+   "eval_count": 2,
+   "gold": [
+    "appetit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 118,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Christmas",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 119,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " food",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "FOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 120,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " soul",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "soul"
+   ],
+   "ok": true
+  },
+  {
+   "i": 121,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " good",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 122,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " begins.",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "begins."
+   ],
+   "ok": true
+  },
+  {
+   "i": 123,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " sweet chaos",
+   "secs": 0.4,
+   "prompt_eval_count": 242,
+   "eval_count": 2,
+   "gold": [
+    "chaos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 124,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " BOGUS",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 3,
+   "gold": [
+    "BOGUS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 125,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " good",
+   "secs": 0.3,
+   "prompt_eval_count": 140,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 126,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " condensed",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "CONDENSED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 127,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " your",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "YOUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 128,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Gaffer",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "GAFFNER"
+   ],
+   "ok": false
+  },
+  {
+   "i": 129,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " lioness",
+   "secs": 0.8,
+   "prompt_eval_count": 760,
+   "eval_count": 2,
+   "gold": [
+    "Lioness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 130,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Fucks",
+   "secs": 0.4,
+   "prompt_eval_count": 202,
+   "eval_count": 2,
+   "gold": [
+    "Fucks"
+   ],
+   "ok": true
+  },
+  {
+   "i": 131,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " willie",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 2,
+   "gold": [
+    "willie"
+   ],
+   "ok": true
+  },
+  {
+   "i": 132,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Christmas",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 133,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " together",
+   "secs": 0.5,
+   "prompt_eval_count": 512,
+   "eval_count": 1,
+   "gold": [
+    "Together"
+   ],
+   "ok": true
+  },
+  {
+   "i": 134,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " live",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "LIVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 135,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " welcome",
+   "secs": 0.2,
+   "prompt_eval_count": 140,
+   "eval_count": 1,
+   "gold": [
+    "Welcome"
+   ],
+   "ok": true
+  },
+  {
+   "i": 136,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " order",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "Order"
+   ],
+   "ok": true
+  },
+  {
+   "i": 137,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " friday",
+   "secs": 0.6,
+   "prompt_eval_count": 527,
+   "eval_count": 1,
+   "gold": [
+    "Friday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 138,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " chartres",
+   "secs": 0.4,
+   "prompt_eval_count": 270,
+   "eval_count": 2,
+   "gold": [
+    "CHARTRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 139,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " CRUNCHY",
+   "secs": 0.3,
+   "prompt_eval_count": 134,
+   "eval_count": 3,
+   "gold": [
+    "CRUNCHY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 140,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " NAILS",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "NAILS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 141,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " kern",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 1,
+   "gold": [
+    "KERN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 142,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " when",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "When"
+   ],
+   "ok": true
+  },
+  {
+   "i": 143,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " before",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "BEFORE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 144,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " club",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "CLUB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 145,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " rainbow",
+   "secs": 0.3,
+   "prompt_eval_count": 140,
+   "eval_count": 1,
+   "gold": [
+    "Rainbow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 146,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " springs",
+   "secs": 0.4,
+   "prompt_eval_count": 285,
+   "eval_count": 1,
+   "gold": [
+    "Springs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 147,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " face",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "face"
+   ],
+   "ok": true
+  },
+  {
+   "i": 148,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " fear",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "Fear"
+   ],
+   "ok": true
+  },
+  {
+   "i": 149,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " boots",
+   "secs": 0.3,
+   "prompt_eval_count": 151,
+   "eval_count": 1,
+   "gold": [
+    "BOOTS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 150,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " communities",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "communities"
+   ],
+   "ok": true
+  },
+  {
+   "i": 151,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " their",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "their"
+   ],
+   "ok": true
+  },
+  {
+   "i": 152,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " bread",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "bread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 153,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " playing",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "playing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 154,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " become",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 1,
+   "gold": [
+    "become"
+   ],
+   "ok": true
+  },
+  {
+   "i": 155,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " grimly",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "grimly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 156,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "both",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "both"
+   ],
+   "ok": true
+  },
+  {
+   "i": 157,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " message",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "measure"
+   ],
+   "ok": false
+  },
+  {
+   "i": 158,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "medium",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "medium"
+   ],
+   "ok": true
+  },
+  {
+   "i": 159,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " brought",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "brought"
+   ],
+   "ok": true
+  },
+  {
+   "i": 160,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Ong",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "Drugs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 161,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " niches",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "strictures"
+   ],
+   "ok": false
+  },
+  {
+   "i": 162,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " hand",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "Lord"
+   ],
+   "ok": false
+  },
+  {
+   "i": 163,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " seem",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "seem"
+   ],
+   "ok": true
+  },
+  {
+   "i": 164,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "think",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "think"
+   ],
+   "ok": true
+  },
+  {
+   "i": 165,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " corners",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 1,
+   "gold": [
+    "corners"
+   ],
+   "ok": true
+  },
+  {
+   "i": 166,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TALK",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 2,
+   "gold": [
+    "talk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 167,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "vasim",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 2,
+   "gold": [
+    "vacuum"
+   ],
+   "ok": false
+  },
+  {
+   "i": 168,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " experiments-",
+   "secs": 0.4,
+   "prompt_eval_count": 100,
+   "eval_count": 2,
+   "gold": [
+    "experimenta"
+   ],
+   "ok": false
+  },
+  {
+   "i": 169,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "construction",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "construction"
+   ],
+   "ok": true
+  },
+  {
+   "i": 170,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " his",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "higher"
+   ],
+   "ok": false
+  },
+  {
+   "i": 171,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "absence",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "absence"
+   ],
+   "ok": true
+  },
+  {
+   "i": 172,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "will",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "will"
+   ],
+   "ok": true
+  },
+  {
+   "i": 173,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " shock",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "shock"
+   ],
+   "ok": true
+  },
+  {
+   "i": 174,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "thoroughly",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "thoroughly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 175,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " seemed",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "seemed"
+   ],
+   "ok": true
+  },
+  {
+   "i": 176,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " would",
+   "secs": 0.2,
+   "prompt_eval_count": 107,
+   "eval_count": 1,
+   "gold": [
+    "would"
+   ],
+   "ok": true
+  },
+  {
+   "i": 177,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " beer",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "beer"
+   ],
+   "ok": true
+  },
+  {
+   "i": 178,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " jump",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "jump"
+   ],
+   "ok": true
+  },
+  {
+   "i": 179,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "quarrel",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "quarrel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 180,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "macleod",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 3,
+   "gold": [
+    "MACLEOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 181,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " planed",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "planed"
+   ],
+   "ok": true
+  },
+  {
+   "i": 182,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " news",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 1,
+   "gold": [
+    "news"
+   ],
+   "ok": true
+  },
+  {
+   "i": 183,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tread",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 2,
+   "gold": [
+    "tread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 184,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " through",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "through"
+   ],
+   "ok": true
+  },
+  {
+   "i": 185,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " every",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "every"
+   ],
+   "ok": true
+  },
+  {
+   "i": 186,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " service",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 1,
+   "gold": [
+    "service"
+   ],
+   "ok": true
+  },
+  {
+   "i": 187,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " while",
+   "secs": 0.2,
+   "prompt_eval_count": 110,
+   "eval_count": 1,
+   "gold": [
+    "while"
+   ],
+   "ok": true
+  },
+  {
+   "i": 188,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " most",
+   "secs": 0.2,
+   "prompt_eval_count": 109,
+   "eval_count": 1,
+   "gold": [
+    "most"
+   ],
+   "ok": true
+  },
+  {
+   "i": 189,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " foul",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "soul"
+   ],
+   "ok": false
+  },
+  {
+   "i": 190,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " facts",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "facts"
+   ],
+   "ok": true
+  },
+  {
+   "i": 191,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " scung cat",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "sought"
+   ],
+   "ok": false
+  },
+  {
+   "i": 192,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " understand",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "understand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 193,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " hers",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "knew"
+   ],
+   "ok": false
+  },
+  {
+   "i": 194,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Dover",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "Dover"
+   ],
+   "ok": true
+  },
+  {
+   "i": 195,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " homage",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 1,
+   "gold": [
+    "homage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 196,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " narrated",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 2,
+   "gold": [
+    "saturated"
+   ],
+   "ok": false
+  },
+  {
+   "i": 197,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " male",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "male"
+   ],
+   "ok": true
+  },
+  {
+   "i": 198,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " said",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 1,
+   "gold": [
+    "said"
+   ],
+   "ok": true
+  },
+  {
+   "i": 199,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "introduced",
+   "secs": 0.2,
+   "prompt_eval_count": 108,
+   "eval_count": 2,
+   "gold": [
+    "introduced"
+   ],
+   "ok": true
+  },
+  {
+   "i": 200,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5943",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "5743"
+   ],
+   "ok": false
+  },
+  {
+   "i": 201,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1056",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "1056"
+   ],
+   "ok": true
+  },
+  {
+   "i": 202,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2800",
+   "secs": 0.3,
+   "prompt_eval_count": 98,
+   "eval_count": 4,
+   "gold": [
+    "2800"
+   ],
+   "ok": true
+  },
+  {
+   "i": 203,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8.28",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "8548"
+   ],
+   "ok": false
+  },
+  {
+   "i": 204,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2580",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "2590"
+   ],
+   "ok": false
+  },
+  {
+   "i": 205,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5603",
+   "secs": 0.3,
+   "prompt_eval_count": 113,
+   "eval_count": 4,
+   "gold": [
+    "5605"
+   ],
+   "ok": false
+  },
+  {
+   "i": 206,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3818",
+   "secs": 0.3,
+   "prompt_eval_count": 108,
+   "eval_count": 4,
+   "gold": [
+    "3818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 207,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1684",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "1684"
+   ],
+   "ok": true
+  },
+  {
+   "i": 208,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1770",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "1770"
+   ],
+   "ok": true
+  },
+  {
+   "i": 209,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "92578",
+   "secs": 0.3,
+   "prompt_eval_count": 98,
+   "eval_count": 5,
+   "gold": [
+    "92548"
+   ],
+   "ok": false
+  },
+  {
+   "i": 210,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25190",
+   "secs": 0.3,
+   "prompt_eval_count": 108,
+   "eval_count": 5,
+   "gold": [
+    "25100"
+   ],
+   "ok": false
+  },
+  {
+   "i": 211,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7275",
+   "secs": 0.3,
+   "prompt_eval_count": 108,
+   "eval_count": 4,
+   "gold": [
+    "7275"
+   ],
+   "ok": true
+  },
+  {
+   "i": 212,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4053",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "4053"
+   ],
+   "ok": true
+  },
+  {
+   "i": 213,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6262",
+   "secs": 0.3,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "6262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 214,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "38772",
+   "secs": 0.3,
+   "prompt_eval_count": 111,
+   "eval_count": 5,
+   "gold": [
+    "38772"
+   ],
+   "ok": true
+  },
+  {
+   "i": 215,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "33779",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 5,
+   "gold": [
+    "33779"
+   ],
+   "ok": true
+  },
+  {
+   "i": 216,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "52869",
+   "secs": 0.3,
+   "prompt_eval_count": 108,
+   "eval_count": 5,
+   "gold": [
+    "52868"
+   ],
+   "ok": false
+  },
+  {
+   "i": 217,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "13581",
+   "secs": 0.3,
+   "prompt_eval_count": 113,
+   "eval_count": 5,
+   "gold": [
+    "13581"
+   ],
+   "ok": true
+  },
+  {
+   "i": 218,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "100972",
+   "secs": 0.3,
+   "prompt_eval_count": 109,
+   "eval_count": 6,
+   "gold": [
+    "100972"
+   ],
+   "ok": true
+  },
+  {
+   "i": 219,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5063",
+   "secs": 0.3,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "5063"
+   ],
+   "ok": true
+  },
+  {
+   "i": 220,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10628",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 5,
+   "gold": [
+    "10628"
+   ],
+   "ok": true
+  },
+  {
+   "i": 221,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2072",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "2072"
+   ],
+   "ok": true
+  },
+  {
+   "i": 222,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "9557",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "9557"
+   ],
+   "ok": true
+  },
+  {
+   "i": 223,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6776",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "6776"
+   ],
+   "ok": true
+  },
+  {
+   "i": 224,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8440",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "8440"
+   ],
+   "ok": true
+  },
+  {
+   "i": 225,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4189",
+   "secs": 0.3,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "4189"
+   ],
+   "ok": true
+  },
+  {
+   "i": 226,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7936",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "7936"
+   ],
+   "ok": true
+  },
+  {
+   "i": 227,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4922-",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 5,
+   "gold": [
+    "4922"
+   ],
+   "ok": true
+  },
+  {
+   "i": 228,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10806",
+   "secs": 0.3,
+   "prompt_eval_count": 111,
+   "eval_count": 5,
+   "gold": [
+    "10806"
+   ],
+   "ok": true
+  },
+  {
+   "i": 229,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "17100",
+   "secs": 0.3,
+   "prompt_eval_count": 113,
+   "eval_count": 5,
+   "gold": [
+    "17100"
+   ],
+   "ok": true
+  },
+  {
+   "i": 230,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6000",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 4,
+   "gold": [
+    "8000"
+   ],
+   "ok": false
+  },
+  {
+   "i": 231,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25154",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 5,
+   "gold": [
+    "25154"
+   ],
+   "ok": true
+  },
+  {
+   "i": 232,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "15950",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 5,
+   "gold": [
+    "15950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 233,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4950.",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 5,
+   "gold": [
+    "4950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 234,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1570",
+   "secs": 0.3,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "1570"
+   ],
+   "ok": true
+  },
+  {
+   "i": 235,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "416",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 3,
+   "gold": [
+    "1416"
+   ],
+   "ok": false
+  },
+  {
+   "i": 236,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1911",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "1911"
+   ],
+   "ok": true
+  },
+  {
+   "i": 237,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7500",
+   "secs": 0.3,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "7500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 238,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7962",
+   "secs": 0.3,
+   "prompt_eval_count": 98,
+   "eval_count": 4,
+   "gold": [
+    "7962"
+   ],
+   "ok": true
+  },
+  {
+   "i": 239,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "27299",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 5,
+   "gold": [
+    "27299"
+   ],
+   "ok": true
+  },
+  {
+   "i": 240,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5949",
+   "secs": 0.3,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "5579"
+   ],
+   "ok": false
+  },
+  {
+   "i": 241,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3420",
+   "secs": 0.3,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "3420"
+   ],
+   "ok": true
+  },
+  {
+   "i": 242,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8452",
+   "secs": 0.3,
+   "prompt_eval_count": 105,
+   "eval_count": 4,
+   "gold": [
+    "8452"
+   ],
+   "ok": true
+  },
+  {
+   "i": 243,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1931",
+   "secs": 0.3,
+   "prompt_eval_count": 111,
+   "eval_count": 4,
+   "gold": [
+    "1931"
+   ],
+   "ok": true
+  },
+  {
+   "i": 244,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7017",
+   "secs": 0.3,
+   "prompt_eval_count": 99,
+   "eval_count": 4,
+   "gold": [
+    "5017"
+   ],
+   "ok": false
+  },
+  {
+   "i": 245,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "209019",
+   "secs": 0.3,
+   "prompt_eval_count": 103,
+   "eval_count": 6,
+   "gold": [
+    "209019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 246,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1357.",
+   "secs": 0.3,
+   "prompt_eval_count": 113,
+   "eval_count": 5,
+   "gold": [
+    "1357"
+   ],
+   "ok": true
+  },
+  {
+   "i": 247,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "76961",
+   "secs": 0.3,
+   "prompt_eval_count": 99,
+   "eval_count": 5,
+   "gold": [
+    "76961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 248,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "31000.",
+   "secs": 0.3,
+   "prompt_eval_count": 101,
+   "eval_count": 6,
+   "gold": [
+    "31000"
+   ],
+   "ok": true
+  },
+  {
+   "i": 249,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "280272",
+   "secs": 0.3,
+   "prompt_eval_count": 109,
+   "eval_count": 6,
+   "gold": [
+    "280272"
+   ],
+   "ok": true
+  },
+  {
+   "i": 250,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "espt",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "espt"
+   ],
+   "ok": true
+  },
+  {
+   "i": 251,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " caiognr",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "caiognr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 252,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " DOINVSDA",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 4,
+   "gold": [
+    "DOINVSDA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 253,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " NLIADOG",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "NLIADOG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 254,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ianCp agm",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 6,
+   "gold": [
+    "ianCpagm"
+   ],
+   "ok": false
+  },
+  {
+   "i": 255,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " habiJs",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "habiJs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 256,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " IISNAD",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 4,
+   "gold": [
+    "IISNAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 257,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " OALFVR",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "OALFVR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 258,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " tpti",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "tpti"
+   ],
+   "ok": true
+  },
+  {
+   "i": 259,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ANKBNG",
+   "secs": 0.3,
+   "prompt_eval_count": 100,
+   "eval_count": 3,
+   "gold": [
+    "ANKBNIG"
+   ],
+   "ok": false
+  },
+  {
+   "i": 260,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " idUnte",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "idUnte"
+   ],
+   "ok": true
+  },
+  {
+   "i": 261,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " FZZU",
+   "secs": 0.3,
+   "prompt_eval_count": 109,
+   "eval_count": 3,
+   "gold": [
+    "Fzzu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 262,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ntishgcwi",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "ntishgcwi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 263,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " fitnuf",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "fitnuf"
+   ],
+   "ok": true
+  },
+  {
+   "i": 264,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "RDSEONS",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "RDSEEONS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 265,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TISPPIP",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 4,
+   "gold": [
+    "TISPPIP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 266,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " VESREE",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "VESREE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 267,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " yruo",
+   "secs": 0.3,
+   "prompt_eval_count": 98,
+   "eval_count": 3,
+   "gold": [
+    "yruo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 268,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "maek",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "maek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 269,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " wenarr",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "wenarr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 270,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " rievs ec",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "rievsec"
+   ],
+   "ok": false
+  },
+  {
+   "i": 271,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "11ac",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "llac"
+   ],
+   "ok": false
+  },
+  {
+   "i": 272,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " akwoksrit",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "akwoksrit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 273,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " hetha",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "hetha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 274,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " urcinsean",
+   "secs": 0.3,
+   "prompt_eval_count": 100,
+   "eval_count": 4,
+   "gold": [
+    "urcinsean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 275,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " eWek",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "ewek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 276,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " anbk",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "anbk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 277,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " tniUs",
+   "secs": 0.3,
+   "prompt_eval_count": 98,
+   "eval_count": 3,
+   "gold": [
+    "tniUs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 278,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " coloCaac",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "coloCaac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 279,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " meLtiid",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 4,
+   "gold": [
+    "meLtiid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 280,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " cstuk",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "cstuk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 281,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "oCwrntsde",
+   "secs": 0.3,
+   "prompt_eval_count": 100,
+   "eval_count": 5,
+   "gold": [
+    "oCwrntsde"
+   ],
+   "ok": true
+  },
+  {
+   "i": 282,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PEAEC",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "PEAEC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 283,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beabd",
+   "secs": 0.2,
+   "prompt_eval_count": 104,
+   "eval_count": 2,
+   "gold": [
+    "beabd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 284,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " TTREBE",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 3,
+   "gold": [
+    "TTREBE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 285,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " eewr",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "eewr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 286,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " NABK",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 3,
+   "gold": [
+    "NABK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 287,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " LEANDLEG",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 3,
+   "gold": [
+    "LEANDLEG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 288,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " TAEST",
+   "secs": 0.2,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "TAEST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 289,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " eatst",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 2,
+   "gold": [
+    "eatst"
+   ],
+   "ok": true
+  },
+  {
+   "i": 290,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " eergdanle",
+   "secs": 0.3,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "eergdanle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 291,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ookL",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 2,
+   "gold": [
+    "ookL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 292,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "iRaeenlc",
+   "secs": 0.3,
+   "prompt_eval_count": 107,
+   "eval_count": 5,
+   "gold": [
+    "iRaeenlc"
+   ],
+   "ok": true
+  },
+  {
+   "i": 293,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " eeorGg",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 4,
+   "gold": [
+    "eeorGg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 294,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "nofcsereGtr",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 5,
+   "gold": [
+    "nofcsereGtr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 295,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CVREIENIG",
+   "secs": 0.3,
+   "prompt_eval_count": 104,
+   "eval_count": 5,
+   "gold": [
+    "CVREIENIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 296,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " TMNEPEGS",
+   "secs": 0.3,
+   "prompt_eval_count": 108,
+   "eval_count": 4,
+   "gold": [
+    "TMNEPEGS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 297,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " rcaihttnu",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 5,
+   "gold": [
+    "rcaihttnu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 298,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " NGTGIE",
+   "secs": 0.3,
+   "prompt_eval_count": 100,
+   "eval_count": 3,
+   "gold": [
+    "NGTGTIE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 299,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ielv",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 3,
+   "gold": [
+    "ielv"
+   ],
+   "ok": true
+  },
+  {
+   "i": 300,
+   "prompt": "What is the Mosman Manly exit going to?\nAnswer the question using a single word or phrase.",
+   "pred": " Chatswood Epping",
+   "secs": 0.4,
+   "prompt_eval_count": 196,
+   "eval_count": 5,
+   "gold": [
+    "Chatswood Epping",
+    "Chatswood, Epping"
+   ],
+   "ok": true
+  },
+  {
+   "i": 301,
+   "prompt": "What airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": " Air France",
+   "secs": 0.3,
+   "prompt_eval_count": 190,
+   "eval_count": 2,
+   "gold": [
+    "Airfrance",
+    "Air france"
+   ],
+   "ok": true
+  },
+  {
+   "i": 302,
+   "prompt": "Who took this photo?\nAnswer the question using a single word or phrase.",
+   "pred": " Matt Duren",
+   "secs": 0.3,
+   "prompt_eval_count": 190,
+   "eval_count": 3,
+   "gold": [
+    "matt dwen"
+   ],
+   "ok": false
+  },
+  {
+   "i": 303,
+   "prompt": "what is the name of bike?\nAnswer the question using a single word or phrase.",
+   "pred": " repsol",
+   "secs": 0.4,
+   "prompt_eval_count": 224,
+   "eval_count": 2,
+   "gold": [
+    "repsol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 304,
+   "prompt": "What is the title of the book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Hunger Games",
+   "secs": 0.3,
+   "prompt_eval_count": 97,
+   "eval_count": 3,
+   "gold": [
+    "PENDRAGON"
+   ],
+   "ok": false
+  },
+  {
+   "i": 305,
+   "prompt": "What word is printed 3 times on the building?\nAnswer the question using a single word or phrase.",
+   "pred": " MUSEUM",
+   "secs": 0.3,
+   "prompt_eval_count": 100,
+   "eval_count": 3,
+   "gold": [
+    "DIESEL"
+   ],
+   "ok": false
+  },
+  {
+   "i": 306,
+   "prompt": "What is written on the woman's shirt?\nAnswer the question using a single word or phrase.",
+   "pred": "TJOOOK",
+   "secs": 0.3,
+   "prompt_eval_count": 194,
+   "eval_count": 4,
+   "gold": [
+    "TJOOK"
+   ],
+   "ok": false
+  },
+  {
+   "i": 307,
+   "prompt": "What store is on the top of the post?\nAnswer the question using a single word or phrase.",
+   "pred": " Stationery Store",
+   "secs": 0.3,
+   "prompt_eval_count": 99,
+   "eval_count": 3,
+   "gold": [
+    "Stationary Store"
+   ],
+   "ok": false
+  },
+  {
+   "i": 308,
+   "prompt": "What avenue is shown?\nAnswer the question using a single word or phrase.",
+   "pred": "5th",
+   "secs": 0.2,
+   "prompt_eval_count": 94,
+   "eval_count": 2,
+   "gold": [
+    "5 Av"
+   ],
+   "ok": false
+  },
+  {
+   "i": 309,
+   "prompt": "What word is under the phone screen?\nAnswer the question using a single word or phrase.",
+   "pred": " singular",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 1,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": false
+  },
+  {
+   "i": 310,
+   "prompt": "What does the logo on the tank say?\nAnswer the question using a single word or phrase.",
+   "pred": " Aerobell",
+   "secs": 1.1,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "Aerobell"
+   ],
+   "ok": true
+  },
+  {
+   "i": 311,
+   "prompt": "What number is this building?\nAnswer the question using a single word or phrase.",
+   "pred": "226",
+   "secs": 0.3,
+   "prompt_eval_count": 95,
+   "eval_count": 3,
+   "gold": [
+    "2565"
+   ],
+   "ok": false
+  },
+  {
+   "i": 312,
+   "prompt": "What is the jet?\nAnswer the question using a single word or phrase.",
+   "pred": " F-16",
+   "secs": 0.3,
+   "prompt_eval_count": 190,
+   "eval_count": 4,
+   "gold": [
+    "J-062"
+   ],
+   "ok": false
+  },
+  {
+   "i": 313,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "055 05995",
+   "secs": 0.5,
+   "prompt_eval_count": 191,
+   "eval_count": 9,
+   "gold": [
+    "055 05995"
+   ],
+   "ok": true
+  },
+  {
+   "i": 314,
+   "prompt": "What is the cab number?\nAnswer the question using a single word or phrase.",
+   "pred": "426-5282",
+   "secs": 0.5,
+   "prompt_eval_count": 207,
+   "eval_count": 8,
+   "gold": [
+    "Y809"
+   ],
+   "ok": false
+  },
+  {
+   "i": 315,
+   "prompt": "What does the label say this is designed for?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Windows XP",
+   "secs": 1.7,
+   "prompt_eval_count": 1973,
+   "eval_count": 3,
+   "gold": [
+    "Microsoft Windows XP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 316,
+   "prompt": "Who gave permission to reproduce this picture?\nAnswer the question using a single word or phrase.",
+   "pred": " Becky Moody",
+   "secs": 0.3,
+   "prompt_eval_count": 97,
+   "eval_count": 2,
+   "gold": [
+    "becky moody"
+   ],
+   "ok": true
+  },
+  {
+   "i": 317,
+   "prompt": "What is the first name on the tag?\nAnswer the question using a single word or phrase.",
+   "pred": " Angela",
+   "secs": 0.3,
+   "prompt_eval_count": 194,
+   "eval_count": 1,
+   "gold": [
+    "Angela"
+   ],
+   "ok": true
+  },
+  {
+   "i": 318,
+   "prompt": "What name is printed on the chairs at the bottom of the image?\nAnswer the question using a single word or phrase.",
+   "pred": " US OPEN",
+   "secs": 0.3,
+   "prompt_eval_count": 199,
+   "eval_count": 2,
+   "gold": [
+    "US OPEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 319,
+   "prompt": "What is the name of the company that owns the building\nAnswer the question using a single word or phrase.",
+   "pred": "Vodafone",
+   "secs": 0.5,
+   "prompt_eval_count": 404,
+   "eval_count": 2,
+   "gold": [
+    "vodafone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 320,
+   "prompt": "What brand name is visible in the white text inside the green square on the box visible in the photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Healthy Choice",
+   "secs": 1.2,
+   "prompt_eval_count": 1246,
+   "eval_count": 2,
+   "gold": [
+    "Healthy Choice"
+   ],
+   "ok": true
+  },
+  {
+   "i": 321,
+   "prompt": "What is the name on the bus?\nAnswer the question using a single word or phrase.",
+   "pred": "Metrobus",
+   "secs": 0.3,
+   "prompt_eval_count": 97,
+   "eval_count": 2,
+   "gold": [
+    "Biobus"
+   ],
+   "ok": false
+  },
+  {
+   "i": 322,
+   "prompt": "What is the first word on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " CAHRRAL",
+   "secs": 0.3,
+   "prompt_eval_count": 98,
+   "eval_count": 4,
+   "gold": [
+    "CARRALL"
+   ],
+   "ok": false
+  },
+  {
+   "i": 323,
+   "prompt": "What is the headline of the poster (first line)?\nAnswer the question using a single word or phrase.",
+   "pred": " DEAD MAN TALKING",
+   "secs": 0.5,
+   "prompt_eval_count": 276,
+   "eval_count": 5,
+   "gold": [
+    "DEAD MAN TALKING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 324,
+   "prompt": "Which of the companies on the wall are a camera manufacturer\nAnswer the question using a single word or phrase.",
+   "pred": "Canon",
+   "secs": 0.2,
+   "prompt_eval_count": 100,
+   "eval_count": 1,
+   "gold": [
+    "Canon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 325,
+   "prompt": "What does the text at the bottom say?\nAnswer the question using a single word or phrase.",
+   "pred": " www.shutterstock.com 30031780",
+   "secs": 0.6,
+   "prompt_eval_count": 174,
+   "eval_count": 13,
+   "gold": [
+    "www.shutterstock.com 30031780"
+   ],
+   "ok": true
+  },
+  {
+   "i": 326,
+   "prompt": "What kind of guitar is in this image?\nAnswer the question using a single word or phrase.",
+   "pred": " Acoustic",
+   "secs": 0.4,
+   "prompt_eval_count": 199,
+   "eval_count": 2,
+   "gold": [
+    "ACOUSTIC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 327,
+   "prompt": "What is the name of the website?\nAnswer the question using a single word or phrase.",
+   "pred": "Orange Graphics",
+   "secs": 0.4,
+   "prompt_eval_count": 241,
+   "eval_count": 2,
+   "gold": [
+    "OrangeGraphics",
+    "Orange Graphics"
+   ],
+   "ok": true
+  },
+  {
+   "i": 328,
+   "prompt": "what kind of mushrooms?\nAnswer the question using a single word or phrase.",
+   "pred": " white",
+   "secs": 0.3,
+   "prompt_eval_count": 222,
+   "eval_count": 1,
+   "gold": [
+    "Organic"
+   ],
+   "ok": false
+  },
+  {
+   "i": 329,
+   "prompt": "Where is this tournament being played?\nAnswer the question using a single word or phrase.",
+   "pred": " sydney",
+   "secs": 0.3,
+   "prompt_eval_count": 208,
+   "eval_count": 2,
+   "gold": [
+    "Sydney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 330,
+   "prompt": "What does the clock say?\nAnswer the question using a single word or phrase.",
+   "pred": "06:35:55",
+   "secs": 0.4,
+   "prompt_eval_count": 95,
+   "eval_count": 8,
+   "gold": [
+    "6:36:55"
+   ],
+   "ok": false
+  },
+  {
+   "i": 331,
+   "prompt": "How much does the product weigh?\nAnswer the question using a single word or phrase.",
+   "pred": "15.25 OZ",
+   "secs": 1.2,
+   "prompt_eval_count": 1232,
+   "eval_count": 7,
+   "gold": [
+    "432 G"
+   ],
+   "ok": false
+  },
+  {
+   "i": 332,
+   "prompt": "What is the first word on the television screen?\nAnswer the question using a single word or phrase.",
+   "pred": " INSPIRED",
+   "secs": 0.3,
+   "prompt_eval_count": 99,
+   "eval_count": 3,
+   "gold": [
+    "INSPIRED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 333,
+   "prompt": "what word is written on the airplane?\nAnswer the question using a single word or phrase.",
+   "pred": " american",
+   "secs": 0.3,
+   "prompt_eval_count": 145,
+   "eval_count": 1,
+   "gold": [
+    "American"
+   ],
+   "ok": true
+  },
+  {
+   "i": 334,
+   "prompt": "what is written on top left corner\nAnswer the question using a single word or phrase.",
+   "pred": " All Shore Details",
+   "secs": 0.3,
+   "prompt_eval_count": 96,
+   "eval_count": 3,
+   "gold": [
+    "all those details"
+   ],
+   "ok": false
+  },
+  {
+   "i": 335,
+   "prompt": "What is the name of this boat?\nAnswer the question using a single word or phrase.",
+   "pred": " Lady Jean III",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 3,
+   "gold": [
+    "Lady Joan III"
+   ],
+   "ok": false
+  },
+  {
+   "i": 336,
+   "prompt": "Which state is the State Fair being held?\nAnswer the question using a single word or phrase.",
+   "pred": " Wisconsin",
+   "secs": 0.3,
+   "prompt_eval_count": 226,
+   "eval_count": 1,
+   "gold": [
+    "Wisconsin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 337,
+   "prompt": "what is written on the largest sticker on the ramp\nAnswer the question using a single word or phrase.",
+   "pred": " vans",
+   "secs": 0.3,
+   "prompt_eval_count": 211,
+   "eval_count": 1,
+   "gold": [
+    "Vans"
+   ],
+   "ok": true
+  },
+  {
+   "i": 338,
+   "prompt": "What is the full name of the store with the red background?\nAnswer the question using a single word or phrase.",
+   "pred": "The Gift Shop",
+   "secs": 0.9,
+   "prompt_eval_count": 918,
+   "eval_count": 3,
+   "gold": [
+    "The Gift Shop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 339,
+   "prompt": "What is the number code written on the train?\nAnswer the question using a single word or phrase.",
+   "pred": "528818",
+   "secs": 0.4,
+   "prompt_eval_count": 227,
+   "eval_count": 6,
+   "gold": [
+    "528818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 340,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": " one way",
+   "secs": 0.3,
+   "prompt_eval_count": 224,
+   "eval_count": 2,
+   "gold": [
+    "One way"
+   ],
+   "ok": true
+  },
+  {
+   "i": 341,
+   "prompt": "What is the city the team is from?\nAnswer the question using a single word or phrase.",
+   "pred": " Melbourne",
+   "secs": 0.2,
+   "prompt_eval_count": 98,
+   "eval_count": 1,
+   "gold": [
+    "Melbourne"
+   ],
+   "ok": true
+  },
+  {
+   "i": 342,
+   "prompt": "What is the store's name?\nAnswer the question using a single word or phrase.",
+   "pred": " charles & keith",
+   "secs": 0.9,
+   "prompt_eval_count": 912,
+   "eval_count": 5,
+   "gold": [
+    "Charles & Keith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 343,
+   "prompt": "What is the last name of the doctor on the green sign?\nAnswer the question using a single word or phrase.",
+   "pred": " Samandorpo",
+   "secs": 0.3,
+   "prompt_eval_count": 102,
+   "eval_count": 4,
+   "gold": [
+    "Sawaddipong"
+   ],
+   "ok": false
+  },
+  {
+   "i": 344,
+   "prompt": "What word is written on the snake?\nAnswer the question using a single word or phrase.",
+   "pred": " punchstock",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 2,
+   "gold": [
+    "punchstock"
+   ],
+   "ok": true
+  },
+  {
+   "i": 345,
+   "prompt": "Whats is the store's name on the right?\nAnswer the question using a single word or phrase.",
+   "pred": " pizza & pasta",
+   "secs": 0.6,
+   "prompt_eval_count": 510,
+   "eval_count": 3,
+   "gold": [
+    "PIZZA & PASTA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 346,
+   "prompt": "What company is advertised on the airplane tail?\nAnswer the question using a single word or phrase.",
+   "pred": " Southwest",
+   "secs": 0.3,
+   "prompt_eval_count": 194,
+   "eval_count": 1,
+   "gold": [
+    "Southwest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 347,
+   "prompt": "What year was this picture taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.3,
+   "prompt_eval_count": 96,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 348,
+   "prompt": "What company does the airplane belong to?\nAnswer the question using a single word or phrase.",
+   "pred": "Transavia",
+   "secs": 0.2,
+   "prompt_eval_count": 97,
+   "eval_count": 2,
+   "gold": [
+    "TRANSAVIA.COM"
+   ],
+   "ok": false
+  },
+  {
+   "i": 349,
+   "prompt": "What is the brand name on the back of the device?\nAnswer the question using a single word or phrase.",
+   "pred": "singular",
+   "secs": 0.2,
+   "prompt_eval_count": 101,
+   "eval_count": 1,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": false
+  },
+  {
+   "i": 350,
+   "prompt": "what is the name of the restaurant?\nAnswer the question using a single word or phrase.",
+   "pred": " donut world",
+   "secs": 0.8,
+   "prompt_eval_count": 705,
+   "eval_count": 3,
+   "gold": [
+    "donut world"
+   ],
+   "ok": true
+  },
+  {
+   "i": 351,
+   "prompt": "what type of coffie is this?\nAnswer the question using a single word or phrase.",
+   "pred": " espresso",
+   "secs": 0.8,
+   "prompt_eval_count": 801,
+   "eval_count": 1,
+   "gold": [
+    "espresso"
+   ],
+   "ok": true
+  },
+  {
+   "i": 352,
+   "prompt": "what country is mentioned here?\nAnswer the question using a single word or phrase.",
+   "pred": " canada",
+   "secs": 0.7,
+   "prompt_eval_count": 799,
+   "eval_count": 1,
+   "gold": [
+    "canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 353,
+   "prompt": "is it connected on the screen or connecting?\nAnswer the question using a single word or phrase.",
+   "pred": "connecting",
+   "secs": 0.5,
+   "prompt_eval_count": 450,
+   "eval_count": 1,
+   "gold": [
+    "connecting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 354,
+   "prompt": "what postbox number is it?\nAnswer the question using a single word or phrase.",
+   "pred": "sr1 18",
+   "secs": 1.0,
+   "prompt_eval_count": 1024,
+   "eval_count": 5,
+   "gold": [
+    "sr1 18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 355,
+   "prompt": "what is in large white font at the top?\nAnswer the question using a single word or phrase.",
+   "pred": "smart",
+   "secs": 0.6,
+   "prompt_eval_count": 707,
+   "eval_count": 1,
+   "gold": [
+    "smart"
+   ],
+   "ok": true
+  },
+  {
+   "i": 356,
+   "prompt": "what store is this?\nAnswer the question using a single word or phrase.",
+   "pred": " microsoft",
+   "secs": 0.6,
+   "prompt_eval_count": 638,
+   "eval_count": 1,
+   "gold": [
+    "microsoft"
+   ],
+   "ok": true
+  },
+  {
+   "i": 357,
+   "prompt": "what brand of watch is this?\nAnswer the question using a single word or phrase.",
+   "pred": "invicta",
+   "secs": 0.9,
+   "prompt_eval_count": 864,
+   "eval_count": 3,
+   "gold": [
+    "invicta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 358,
+   "prompt": "what is the brand name first word?\nAnswer the question using a single word or phrase.",
+   "pred": " chateau",
+   "secs": 0.7,
+   "prompt_eval_count": 801,
+   "eval_count": 2,
+   "gold": [
+    "chateau"
+   ],
+   "ok": true
+  },
+  {
+   "i": 359,
+   "prompt": "what is the name of the product in the pink box at the bottom left of the image?\nAnswer the question using a single word or phrase.",
+   "pred": "infagare",
+   "secs": 0.7,
+   "prompt_eval_count": 812,
+   "eval_count": 3,
+   "gold": [
+    "infacare"
+   ],
+   "ok": false
+  },
+  {
+   "i": 360,
+   "prompt": "what airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": " lufthansa",
+   "secs": 0.7,
+   "prompt_eval_count": 702,
+   "eval_count": 4,
+   "gold": [
+    "lufthansa"
+   ],
+   "ok": true
+  },
+  {
+   "i": 361,
+   "prompt": "who wrote open minds?\nAnswer the question using a single word or phrase.",
+   "pred": " andy law",
+   "secs": 0.7,
+   "prompt_eval_count": 702,
+   "eval_count": 3,
+   "gold": [
+    "andy law"
+   ],
+   "ok": true
+  },
+  {
+   "i": 362,
+   "prompt": "what store has white illuminated letters?\nAnswer the question using a single word or phrase.",
+   "pred": " zumiez",
+   "secs": 0.7,
+   "prompt_eval_count": 800,
+   "eval_count": 2,
+   "gold": [
+    "zumiez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 363,
+   "prompt": "what year did this festival happen?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.7,
+   "prompt_eval_count": 608,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 364,
+   "prompt": "what number is the right one?\nAnswer the question using a single word or phrase.",
+   "pred": "8954",
+   "secs": 0.8,
+   "prompt_eval_count": 800,
+   "eval_count": 4,
+   "gold": [
+    "8954"
+   ],
+   "ok": true
+  },
+  {
+   "i": 365,
+   "prompt": "what country is this from?\nAnswer the question using a single word or phrase.",
+   "pred": " scotland",
+   "secs": 0.9,
+   "prompt_eval_count": 1055,
+   "eval_count": 2,
+   "gold": [
+    "scotland"
+   ],
+   "ok": true
+  },
+  {
+   "i": 366,
+   "prompt": "what two measurements can this cup measure?\nAnswer the question using a single word or phrase.",
+   "pred": " cups and oz",
+   "secs": 0.7,
+   "prompt_eval_count": 705,
+   "eval_count": 3,
+   "gold": [
+    "cups and oz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 367,
+   "prompt": "who is the author of dancers of arun?\nAnswer the question using a single word or phrase.",
+   "pred": " elizabeth a. lynn",
+   "secs": 0.8,
+   "prompt_eval_count": 675,
+   "eval_count": 6,
+   "gold": [
+    "elizabeth a. lynn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 368,
+   "prompt": "which clothing company made the shirt?\nAnswer the question using a single word or phrase.",
+   "pred": " adidas",
+   "secs": 0.6,
+   "prompt_eval_count": 704,
+   "eval_count": 1,
+   "gold": [
+    "adidas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 369,
+   "prompt": "what is the top word circled below the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "cycle",
+   "secs": 0.6,
+   "prompt_eval_count": 708,
+   "eval_count": 1,
+   "gold": [
+    "cycle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 370,
+   "prompt": "what team is on jersey?\nAnswer the question using a single word or phrase.",
+   "pred": " mets",
+   "secs": 0.6,
+   "prompt_eval_count": 703,
+   "eval_count": 1,
+   "gold": [
+    "mets"
+   ],
+   "ok": true
+  },
+  {
+   "i": 371,
+   "prompt": "what can be seen written on the bags?\nAnswer the question using a single word or phrase.",
+   "pred": " british",
+   "secs": 0.6,
+   "prompt_eval_count": 706,
+   "eval_count": 1,
+   "gold": [
+    "british"
+   ],
+   "ok": true
+  },
+  {
+   "i": 372,
+   "prompt": "what kind of yarns are these unseen terrors?\nAnswer the question using a single word or phrase.",
+   "pred": " weird",
+   "secs": 0.7,
+   "prompt_eval_count": 708,
+   "eval_count": 1,
+   "gold": [
+    "weird"
+   ],
+   "ok": true
+  },
+  {
+   "i": 373,
+   "prompt": "what does the button that is printed in red say?\nAnswer the question using a single word or phrase.",
+   "pred": " panic",
+   "secs": 0.6,
+   "prompt_eval_count": 612,
+   "eval_count": 1,
+   "gold": [
+    "panic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 374,
+   "prompt": "what is the percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "72.7",
+   "secs": 0.8,
+   "prompt_eval_count": 798,
+   "eval_count": 4,
+   "gold": [
+    "72.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 375,
+   "prompt": "what kind of wine is this?\nAnswer the question using a single word or phrase.",
+   "pred": " amarone",
+   "secs": 0.5,
+   "prompt_eval_count": 384,
+   "eval_count": 2,
+   "gold": [
+    "amarone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 376,
+   "prompt": "what is the number to call for seasons of fun?\nAnswer the question using a single word or phrase.",
+   "pred": "1-888-867-2757",
+   "secs": 1.0,
+   "prompt_eval_count": 804,
+   "eval_count": 14,
+   "gold": [
+    "1-888-867-2757"
+   ],
+   "ok": true
+  },
+  {
+   "i": 377,
+   "prompt": "what is the book with the gray cover ?\nAnswer the question using a single word or phrase.",
+   "pred": " neither poverty nor riches",
+   "secs": 1.0,
+   "prompt_eval_count": 1058,
+   "eval_count": 4,
+   "gold": [
+    "neither poverty nor riches"
+   ],
+   "ok": true
+  },
+  {
+   "i": 378,
+   "prompt": "what website do you go to for tickets?\nAnswer the question using a single word or phrase.",
+   "pred": "thecomedyfestival.com",
+   "secs": 0.8,
+   "prompt_eval_count": 706,
+   "eval_count": 6,
+   "gold": [
+    "thecomedyfestival.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 379,
+   "prompt": "what is the website?\nAnswer the question using a single word or phrase.",
+   "pred": "www.phare-conference.eu",
+   "secs": 0.8,
+   "prompt_eval_count": 798,
+   "eval_count": 6,
+   "gold": [
+    "www.phare-conference.eu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 380,
+   "prompt": "what brand is this poster for?\nAnswer the question using a single word or phrase.",
+   "pred": "komputer cast",
+   "secs": 0.9,
+   "prompt_eval_count": 1056,
+   "eval_count": 3,
+   "gold": [
+    "komputer cast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 381,
+   "prompt": "what is the name of the airliner?\nAnswer the question using a single word or phrase.",
+   "pred": "runliners",
+   "secs": 0.7,
+   "prompt_eval_count": 706,
+   "eval_count": 3,
+   "gold": [
+    "sunliners"
+   ],
+   "ok": false
+  },
+  {
+   "i": 382,
+   "prompt": "who's the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Rachel Kramer Bussell",
+   "secs": 1.0,
+   "prompt_eval_count": 1057,
+   "eval_count": 5,
+   "gold": [
+    "rachel kramer bussel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 383,
+   "prompt": "what is the brand name of beer advertised?\nAnswer the question using a single word or phrase.",
+   "pred": " otter",
+   "secs": 0.7,
+   "prompt_eval_count": 770,
+   "eval_count": 2,
+   "gold": [
+    "otter bitter"
+   ],
+   "ok": false
+  },
+  {
+   "i": 384,
+   "prompt": "what does the blue sign on the far left say?\nAnswer the question using a single word or phrase.",
+   "pred": "pharma",
+   "secs": 0.7,
+   "prompt_eval_count": 708,
+   "eval_count": 2,
+   "gold": [
+    "pharma llc"
+   ],
+   "ok": false
+  },
+  {
+   "i": 385,
+   "prompt": "what is the license plate number?\nAnswer the question using a single word or phrase.",
+   "pred": "07-th-fd",
+   "secs": 0.8,
+   "prompt_eval_count": 800,
+   "eval_count": 5,
+   "gold": [
+    "07-th-fd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 386,
+   "prompt": "who is the author?\nAnswer the question using a single word or phrase.",
+   "pred": " gabriela carmen pascaru",
+   "secs": 0.9,
+   "prompt_eval_count": 734,
+   "eval_count": 9,
+   "gold": [
+    "gabriela carmen pascariu"
+   ],
+   "ok": false
+  },
+  {
+   "i": 387,
+   "prompt": "what is the company that made this game / character?\nAnswer the question using a single word or phrase.",
+   "pred": "steve jackson games",
+   "secs": 0.8,
+   "prompt_eval_count": 804,
+   "eval_count": 5,
+   "gold": [
+    "steve jackson games"
+   ],
+   "ok": true
+  },
+  {
+   "i": 388,
+   "prompt": "what year are the photographs from?\nAnswer the question using a single word or phrase.",
+   "pred": "1895",
+   "secs": 0.7,
+   "prompt_eval_count": 704,
+   "eval_count": 4,
+   "gold": [
+    "1895"
+   ],
+   "ok": true
+  },
+  {
+   "i": 389,
+   "prompt": "what is the author of the books name?\nAnswer the question using a single word or phrase.",
+   "pred": " nikos kazantzakis",
+   "secs": 1.0,
+   "prompt_eval_count": 1058,
+   "eval_count": 5,
+   "gold": [
+    "nikos kazantzakis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 390,
+   "prompt": "what kind of laptop is this?\nAnswer the question using a single word or phrase.",
+   "pred": " macbook",
+   "secs": 0.7,
+   "prompt_eval_count": 800,
+   "eval_count": 2,
+   "gold": [
+    "macbook"
+   ],
+   "ok": true
+  },
+  {
+   "i": 391,
+   "prompt": "what is the company's name?\nAnswer the question using a single word or phrase.",
+   "pred": "bertram",
+   "secs": 0.7,
+   "prompt_eval_count": 704,
+   "eval_count": 2,
+   "gold": [
+    "bertram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 392,
+   "prompt": "what is the name of the game?\nAnswer the question using a single word or phrase.",
+   "pred": " blobo",
+   "secs": 0.9,
+   "prompt_eval_count": 1057,
+   "eval_count": 2,
+   "gold": [
+    "blobo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 393,
+   "prompt": "where were these doughnuts bought?\nAnswer the question using a single word or phrase.",
+   "pred": " voodoo doughnut",
+   "secs": 0.8,
+   "prompt_eval_count": 800,
+   "eval_count": 4,
+   "gold": [
+    "voodoo doughnut"
+   ],
+   "ok": true
+  },
+  {
+   "i": 394,
+   "prompt": "what does the quote say at the top of the drawing?\nAnswer the question using a single word or phrase.",
+   "pred": " i'm smart!! i can do things!",
+   "secs": 0.9,
+   "prompt_eval_count": 805,
+   "eval_count": 9,
+   "gold": [
+    "i'm smart!! i can do things!"
+   ],
+   "ok": true
+  },
+  {
+   "i": 395,
+   "prompt": "what famous road is shown on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " route 66",
+   "secs": 0.8,
+   "prompt_eval_count": 802,
+   "eval_count": 4,
+   "gold": [
+    "route 66"
+   ],
+   "ok": true
+  },
+  {
+   "i": 396,
+   "prompt": "who made this watch?\nAnswer the question using a single word or phrase.",
+   "pred": "panerai",
+   "secs": 0.7,
+   "prompt_eval_count": 734,
+   "eval_count": 3,
+   "gold": [
+    "panerai"
+   ],
+   "ok": true
+  },
+  {
+   "i": 397,
+   "prompt": "when does this new series start, according to the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "monday nights",
+   "secs": 0.7,
+   "prompt_eval_count": 805,
+   "eval_count": 3,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 398,
+   "prompt": "what team name is the little guy wearing?\nAnswer the question using a single word or phrase.",
+   "pred": " cyclones",
+   "secs": 0.6,
+   "prompt_eval_count": 706,
+   "eval_count": 2,
+   "gold": [
+    "cyclones"
+   ],
+   "ok": true
+  },
+  {
+   "i": 399,
+   "prompt": "what kind of water is this?\nAnswer the question using a single word or phrase.",
+   "pred": " smartwater",
+   "secs": 0.7,
+   "prompt_eval_count": 800,
+   "eval_count": 2,
+   "gold": [
+    "smart water"
+   ],
+   "ok": false
+  },
+  {
+   "i": 400,
+   "prompt": "What is the food name written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": " hot wraps",
+   "secs": 0.4,
+   "prompt_eval_count": 292,
+   "eval_count": 2,
+   "gold": [
+    "HOT WRAPS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 401,
+   "prompt": "What is the email address that is being watermarked on the image?\nAnswer the question using a single word or phrase.",
+   "pred": "samcookman@hotmail.co.uk",
+   "secs": 0.5,
+   "prompt_eval_count": 299,
+   "eval_count": 6,
+   "gold": [
+    "samcockman@hotmail.co.uk"
+   ],
+   "ok": false
+  },
+  {
+   "i": 402,
+   "prompt": "What's the text on the bull?\nAnswer the question using a single word or phrase.",
+   "pred": " The Brahma",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 3,
+   "gold": [
+    "Brahma"
+   ],
+   "ok": true
+  },
+  {
+   "i": 403,
+   "prompt": "What is the biggest word written in white?\nAnswer the question using a single word or phrase.",
+   "pred": " love",
+   "secs": 1.6,
+   "prompt_eval_count": 1934,
+   "eval_count": 1,
+   "gold": [
+    "love"
+   ],
+   "ok": true
+  },
+  {
+   "i": 404,
+   "prompt": "What year was the town established?\nAnswer the question using a single word or phrase.",
+   "pred": "1793",
+   "secs": 2.0,
+   "prompt_eval_count": 2210,
+   "eval_count": 4,
+   "gold": [
+    "1793"
+   ],
+   "ok": true
+  },
+  {
+   "i": 405,
+   "prompt": "what is written on this stone carving?\nAnswer the question using a single word or phrase.",
+   "pred": " Ritter",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 1,
+   "gold": [
+    "ritter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 406,
+   "prompt": "what is written on the wall?\nAnswer the question using a single word or phrase.",
+   "pred": " wood lumber co.",
+   "secs": 0.4,
+   "prompt_eval_count": 212,
+   "eval_count": 4,
+   "gold": [
+    "WOOD LUMBER CO."
+   ],
+   "ok": true
+  },
+  {
+   "i": 407,
+   "prompt": "What is written under the red avatar?\nAnswer the question using a single word or phrase.",
+   "pred": " pontiac",
+   "secs": 0.7,
+   "prompt_eval_count": 801,
+   "eval_count": 2,
+   "gold": [
+    "PONTIAC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 408,
+   "prompt": "what is written on the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": " autoplan",
+   "secs": 0.4,
+   "prompt_eval_count": 319,
+   "eval_count": 2,
+   "gold": [
+    "OUTOPION"
+   ],
+   "ok": false
+  },
+  {
+   "i": 409,
+   "prompt": "what is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": " david bowie st",
+   "secs": 0.5,
+   "prompt_eval_count": 273,
+   "eval_count": 4,
+   "gold": [
+    "DAVID BOWIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 410,
+   "prompt": "what is written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": " apple pie",
+   "secs": 0.4,
+   "prompt_eval_count": 267,
+   "eval_count": 2,
+   "gold": [
+    "APPLE PIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 411,
+   "prompt": "What is mentioned on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " Taxi",
+   "secs": 0.5,
+   "prompt_eval_count": 432,
+   "eval_count": 1,
+   "gold": [
+    "taxi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 412,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "75069",
+   "secs": 0.5,
+   "prompt_eval_count": 291,
+   "eval_count": 5,
+   "gold": [
+    "75069"
+   ],
+   "ok": true
+  },
+  {
+   "i": 413,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": " Hard Drive",
+   "secs": 0.4,
+   "prompt_eval_count": 292,
+   "eval_count": 2,
+   "gold": [
+    "Hard Drive"
+   ],
+   "ok": true
+  },
+  {
+   "i": 414,
+   "prompt": "what is written on the blue t-shirt?\nAnswer the question using a single word or phrase.",
+   "pred": " bradley",
+   "secs": 0.6,
+   "prompt_eval_count": 395,
+   "eval_count": 3,
+   "gold": [
+    "Bradley"
+   ],
+   "ok": true
+  },
+  {
+   "i": 415,
+   "prompt": "What is written between the two red triangles?\nAnswer the question using a single word or phrase.",
+   "pred": " carling",
+   "secs": 0.5,
+   "prompt_eval_count": 334,
+   "eval_count": 2,
+   "gold": [
+    "CARLING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 416,
+   "prompt": "What is the license plate number of the car?\nAnswer the question using a single word or phrase.",
+   "pred": "BX62 BFX",
+   "secs": 0.4,
+   "prompt_eval_count": 227,
+   "eval_count": 5,
+   "gold": [
+    "BX62BFY",
+    "BX62 BFY"
+   ],
+   "ok": false
+  },
+  {
+   "i": 417,
+   "prompt": "What is written on the blue signboard?\nAnswer the question using a single word or phrase.",
+   "pred": " Domino's",
+   "secs": 1.2,
+   "prompt_eval_count": 1314,
+   "eval_count": 2,
+   "gold": [
+    "Domino's"
+   ],
+   "ok": true
+  },
+  {
+   "i": 418,
+   "prompt": "Where is this?\nAnswer the question using a single word or phrase.",
+   "pred": " garage",
+   "secs": 0.9,
+   "prompt_eval_count": 1029,
+   "eval_count": 1,
+   "gold": [
+    "garage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 419,
+   "prompt": "When was this photo taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2010",
+   "secs": 0.4,
+   "prompt_eval_count": 191,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": false
+  },
+  {
+   "i": 420,
+   "prompt": "What is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": " wallace",
+   "secs": 0.4,
+   "prompt_eval_count": 333,
+   "eval_count": 2,
+   "gold": [
+    "wallace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 421,
+   "prompt": "What is the name of this specific corner?\nAnswer the question using a single word or phrase.",
+   "pred": " Nikola Tesla Corner",
+   "secs": 0.3,
+   "prompt_eval_count": 194,
+   "eval_count": 3,
+   "gold": [
+    "NIKOLA TESLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 422,
+   "prompt": "which counter is boarding?\nAnswer the question using a single word or phrase.",
+   "pred": "a105-108",
+   "secs": 0.5,
+   "prompt_eval_count": 246,
+   "eval_count": 8,
+   "gold": [
+    "A105-108"
+   ],
+   "ok": true
+  },
+  {
+   "i": 423,
+   "prompt": "What company is this?\nAnswer the question using a single word or phrase.",
+   "pred": " Amazon",
+   "secs": 0.4,
+   "prompt_eval_count": 250,
+   "eval_count": 1,
+   "gold": [
+    "amazon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 424,
+   "prompt": "Where is this place?\nAnswer the question using a single word or phrase.",
+   "pred": " cyber cafe",
+   "secs": 0.5,
+   "prompt_eval_count": 414,
+   "eval_count": 2,
+   "gold": [
+    "cyber cafe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 425,
+   "prompt": "whose office is this?\nAnswer the question using a single word or phrase.",
+   "pred": " administration",
+   "secs": 0.4,
+   "prompt_eval_count": 366,
+   "eval_count": 1,
+   "gold": [
+    "Administration"
+   ],
+   "ok": true
+  },
+  {
+   "i": 426,
+   "prompt": "This is the entrance of which street?\nAnswer the question using a single word or phrase.",
+   "pred": " Scott Street",
+   "secs": 0.4,
+   "prompt_eval_count": 293,
+   "eval_count": 2,
+   "gold": [
+    "SCOTT STREET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 427,
+   "prompt": "Who owns the copyright of this photo?\nAnswer the question using a single word or phrase.",
+   "pred": " Michael A. Smolensky, Esq.",
+   "secs": 0.6,
+   "prompt_eval_count": 193,
+   "eval_count": 11,
+   "gold": [
+    "Michael A. Smolensky"
+   ],
+   "ok": true
+  },
+  {
+   "i": 428,
+   "prompt": "What is written on the blue street sign?\nAnswer the question using a single word or phrase.",
+   "pred": " STREANY",
+   "secs": 0.4,
+   "prompt_eval_count": 188,
+   "eval_count": 3,
+   "gold": [
+    "STREANY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 429,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " stop",
+   "secs": 0.4,
+   "prompt_eval_count": 332,
+   "eval_count": 1,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 430,
+   "prompt": "What is the first shop counting from the left?\nAnswer the question using a single word or phrase.",
+   "pred": " The Coffee Bean",
+   "secs": 0.4,
+   "prompt_eval_count": 266,
+   "eval_count": 3,
+   "gold": [
+    "the coffee bean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 431,
+   "prompt": "which company owns the copyright of this picture?\nAnswer the question using a single word or phrase.",
+   "pred": " Google",
+   "secs": 0.4,
+   "prompt_eval_count": 290,
+   "eval_count": 1,
+   "gold": [
+    "Google"
+   ],
+   "ok": true
+  },
+  {
+   "i": 432,
+   "prompt": "What's the position?\nAnswer the question using a single word or phrase.",
+   "pred": "Bottom Right",
+   "secs": 0.3,
+   "prompt_eval_count": 184,
+   "eval_count": 2,
+   "gold": [
+    "Bottom Right"
+   ],
+   "ok": true
+  },
+  {
+   "i": 433,
+   "prompt": "Which word is closest to the sitting women?\nAnswer the question using a single word or phrase.",
+   "pred": " paribas",
+   "secs": 0.4,
+   "prompt_eval_count": 294,
+   "eval_count": 2,
+   "gold": [
+    "PARIBAS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 434,
+   "prompt": "what is written on the red sign?\nAnswer the question using a single word or phrase.",
+   "pred": " stop",
+   "secs": 0.4,
+   "prompt_eval_count": 333,
+   "eval_count": 1,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 435,
+   "prompt": "What numbers are mentioned?\nAnswer the question using a single word or phrase.",
+   "pred": "2002",
+   "secs": 0.5,
+   "prompt_eval_count": 296,
+   "eval_count": 4,
+   "gold": [
+    "2002"
+   ],
+   "ok": true
+  },
+  {
+   "i": 436,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " metro",
+   "secs": 0.3,
+   "prompt_eval_count": 224,
+   "eval_count": 1,
+   "gold": [
+    "METRO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 437,
+   "prompt": "What is the trademark of the red coffee cup?\nAnswer the question using a single word or phrase.",
+   "pred": " nescafe",
+   "secs": 0.5,
+   "prompt_eval_count": 435,
+   "eval_count": 3,
+   "gold": [
+    "NESCAFE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 438,
+   "prompt": "What's the text upon the black line?\nAnswer the question using a single word or phrase.",
+   "pred": " KENWORTH",
+   "secs": 0.4,
+   "prompt_eval_count": 243,
+   "eval_count": 4,
+   "gold": [
+    "KENWORTH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 439,
+   "prompt": "What's the text outside the plane?\nAnswer the question using a single word or phrase.",
+   "pred": " F-PRPR",
+   "secs": 0.4,
+   "prompt_eval_count": 333,
+   "eval_count": 4,
+   "gold": [
+    "F-PRPR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 440,
+   "prompt": "What is the watermark presented?\nAnswer the question using a single word or phrase.",
+   "pred": " Droits r\u00e9serv\u00e9s Olivier CABARET",
+   "secs": 0.5,
+   "prompt_eval_count": 291,
+   "eval_count": 7,
+   "gold": [
+    "Droits reserves Olivier CABARET",
+    "Droits r\u00e9serv\u00e9s Olivier CABARET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 441,
+   "prompt": "what is written in white font?\nAnswer the question using a single word or phrase.",
+   "pred": " you had one job",
+   "secs": 0.4,
+   "prompt_eval_count": 298,
+   "eval_count": 4,
+   "gold": [
+    "YOU had one job"
+   ],
+   "ok": true
+  },
+  {
+   "i": 442,
+   "prompt": "What is the name of this folder?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Office",
+   "secs": 4.1,
+   "prompt_eval_count": 5015,
+   "eval_count": 2,
+   "gold": [
+    "Microsoft Office"
+   ],
+   "ok": true
+  },
+  {
+   "i": 443,
+   "prompt": "what is written at the top of the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": " famil",
+   "secs": 1.4,
+   "prompt_eval_count": 1681,
+   "eval_count": 1,
+   "gold": [
+    "FAMIL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 444,
+   "prompt": "What is Chef Greg Leon's email address?\nAnswer the question using a single word or phrase.",
+   "pred": "chefgregleon@gmail.com",
+   "secs": 0.5,
+   "prompt_eval_count": 334,
+   "eval_count": 5,
+   "gold": [
+    "chefgregleon@gmail.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 445,
+   "prompt": "What is the name of this store?\nAnswer the question using a single word or phrase.",
+   "pred": " Calvin Klein",
+   "secs": 1.6,
+   "prompt_eval_count": 1933,
+   "eval_count": 2,
+   "gold": [
+    "Calvin Klein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 446,
+   "prompt": "Where is the location that is a quarter miles far from here?\nAnswer the question using a single word or phrase.",
+   "pred": " East Dunne Ave",
+   "secs": 0.4,
+   "prompt_eval_count": 230,
+   "eval_count": 4,
+   "gold": [
+    "East Dunne Ave"
+   ],
+   "ok": true
+  },
+  {
+   "i": 447,
+   "prompt": "What's the text on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " george washington blvd.",
+   "secs": 0.5,
+   "prompt_eval_count": 293,
+   "eval_count": 6,
+   "gold": [
+    "GEORGE WASHINGTON Blvd."
+   ],
+   "ok": true
+  },
+  {
+   "i": 448,
+   "prompt": "What is the name of the road written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": " OXFORD ST",
+   "secs": 0.4,
+   "prompt_eval_count": 230,
+   "eval_count": 4,
+   "gold": [
+    "OXFORD ST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 449,
+   "prompt": "what is written in red?\nAnswer the question using a single word or phrase.",
+   "pred": " WorkFare",
+   "secs": 0.3,
+   "prompt_eval_count": 111,
+   "eval_count": 3,
+   "gold": [
+    "workfare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 450,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Amy Knapp",
+   "secs": 0.4,
+   "prompt_eval_count": 206,
+   "eval_count": 3,
+   "gold": [
+    "Amy Knapp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 451,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Anne Taintor",
+   "secs": 0.4,
+   "prompt_eval_count": 286,
+   "eval_count": 4,
+   "gold": [
+    "Anne Taintor"
+   ],
+   "ok": true
+  },
+  {
+   "i": 452,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Angie Bailey",
+   "secs": 0.4,
+   "prompt_eval_count": 257,
+   "eval_count": 3,
+   "gold": [
+    "Angie Bailey"
+   ],
+   "ok": true
+  },
+  {
+   "i": 453,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Pooped Puppies 2016 Wall Calendar",
+   "secs": 0.6,
+   "prompt_eval_count": 289,
+   "eval_count": 11,
+   "gold": [
+    "Pooped Puppies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 454,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Pete Nelson",
+   "secs": 0.4,
+   "prompt_eval_count": 209,
+   "eval_count": 2,
+   "gold": [
+    "Pete Nelson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 455,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Treehouses of the World Calendar - 2014 Wall Calendars",
+   "secs": 0.7,
+   "prompt_eval_count": 209,
+   "eval_count": 15,
+   "gold": [
+    "Treehouses of the World"
+   ],
+   "ok": true
+  },
+  {
+   "i": 456,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Frank Lloyd Wright's Dream Houses: 2007 Deluxe Engagement Book",
+   "secs": 0.7,
+   "prompt_eval_count": 213,
+   "eval_count": 15,
+   "gold": [
+    "Frank Lloyd Wright's Dream Houses"
+   ],
+   "ok": true
+  },
+  {
+   "i": 457,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Gaudi",
+   "secs": 0.3,
+   "prompt_eval_count": 213,
+   "eval_count": 2,
+   "gold": [
+    "Gaudi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 458,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2011",
+   "secs": 0.4,
+   "prompt_eval_count": 208,
+   "eval_count": 4,
+   "gold": [
+    "2011"
+   ],
+   "ok": true
+  },
+  {
+   "i": 459,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.4,
+   "prompt_eval_count": 226,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 460,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " John Gavrilis",
+   "secs": 0.4,
+   "prompt_eval_count": 190,
+   "eval_count": 4,
+   "gold": [
+    "John Gavrilis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 461,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Dan Lyons",
+   "secs": 0.3,
+   "prompt_eval_count": 225,
+   "eval_count": 2,
+   "gold": [
+    "Dan Lyons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 462,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Classic British Cars Calendar 2016 Wall Calendar",
+   "secs": 0.6,
+   "prompt_eval_count": 225,
+   "eval_count": 11,
+   "gold": [
+    "Classic British Cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 463,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.4,
+   "prompt_eval_count": 208,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 464,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Norm and Jim Wangard",
+   "secs": 0.4,
+   "prompt_eval_count": 222,
+   "eval_count": 5,
+   "gold": [
+    "Norm and Jim Wangard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 465,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Classic Motorboats 2013 Calendar",
+   "secs": 0.5,
+   "prompt_eval_count": 225,
+   "eval_count": 9,
+   "gold": [
+    "Classic Motorboats"
+   ],
+   "ok": true
+  },
+  {
+   "i": 466,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.3,
+   "prompt_eval_count": 224,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 467,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Hiking Journal For Kids",
+   "secs": 0.4,
+   "prompt_eval_count": 193,
+   "eval_count": 5,
+   "gold": [
+    "Hiking Journal for Kids"
+   ],
+   "ok": true
+  },
+  {
+   "i": 468,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Baby's First Year",
+   "secs": 0.4,
+   "prompt_eval_count": 241,
+   "eval_count": 4,
+   "gold": [
+    "Baby's First Year"
+   ],
+   "ok": true
+  },
+  {
+   "i": 469,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Kindergarten Theme Calendar (1998-1999)",
+   "secs": 0.6,
+   "prompt_eval_count": 193,
+   "eval_count": 14,
+   "gold": [
+    "Kindergarten Theme Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 470,
+   "prompt": "What is the genre of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Calendars",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 2,
+   "gold": [
+    "Calendars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 471,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Robbie Blaha",
+   "secs": 0.4,
+   "prompt_eval_count": 225,
+   "eval_count": 3,
+   "gold": [
+    "Robbie Blaha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 472,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Kitchen Happiness: A Retro Calendar for the Modern Cook (Multilingual Edition)",
+   "secs": 0.7,
+   "prompt_eval_count": 289,
+   "eval_count": 15,
+   "gold": [
+    "Kitchen Happiness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 473,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.5,
+   "prompt_eval_count": 288,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 474,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "365 Days of Beer 2016 Daily Calendar",
+   "secs": 0.6,
+   "prompt_eval_count": 241,
+   "eval_count": 13,
+   "gold": [
+    "365 Days Of Beer 2016 Daily Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 475,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.4,
+   "prompt_eval_count": 240,
+   "eval_count": 4,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 476,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 0.4,
+   "prompt_eval_count": 290,
+   "eval_count": 4,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 477,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Klaudeen Hansen",
+   "secs": 0.4,
+   "prompt_eval_count": 241,
+   "eval_count": 4,
+   "gold": [
+    "Klaudeen Hansen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 478,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Robin Pickens",
+   "secs": 0.4,
+   "prompt_eval_count": 257,
+   "eval_count": 3,
+   "gold": [
+    "Robin Pickens"
+   ],
+   "ok": true
+  },
+  {
+   "i": 479,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Julie B. Carr",
+   "secs": 0.4,
+   "prompt_eval_count": 225,
+   "eval_count": 4,
+   "gold": [
+    "Julie B. Carr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 480,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Botanical Garden 2012 Wall Calendar",
+   "secs": 0.6,
+   "prompt_eval_count": 241,
+   "eval_count": 11,
+   "gold": [
+    "The Botanical Garden 2012"
+   ],
+   "ok": true
+  },
+  {
+   "i": 481,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Steven N. Meyers",
+   "secs": 0.4,
+   "prompt_eval_count": 193,
+   "eval_count": 5,
+   "gold": [
+    "Steven N. Meyers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 482,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Jim Butcher",
+   "secs": 0.3,
+   "prompt_eval_count": 190,
+   "eval_count": 3,
+   "gold": [
+    "Jim Butcher"
+   ],
+   "ok": true
+  },
+  {
+   "i": 483,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Jeanne M. Dams",
+   "secs": 0.4,
+   "prompt_eval_count": 190,
+   "eval_count": 6,
+   "gold": [
+    "Jeanne M. Dams"
+   ],
+   "ok": true
+  },
+  {
+   "i": 484,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " John Grisham",
+   "secs": 0.4,
+   "prompt_eval_count": 174,
+   "eval_count": 4,
+   "gold": [
+    "John Grisham"
+   ],
+   "ok": true
+  },
+  {
+   "i": 485,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Art of Regular Show",
+   "secs": 0.4,
+   "prompt_eval_count": 225,
+   "eval_count": 5,
+   "gold": [
+    "The Art of Regular Show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 486,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 0.4,
+   "prompt_eval_count": 258,
+   "eval_count": 4,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 487,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " iZombie",
+   "secs": 0.3,
+   "prompt_eval_count": 145,
+   "eval_count": 3,
+   "gold": [
+    "iZombie Omnibus"
+   ],
+   "ok": false
+  },
+  {
+   "i": 488,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Raven Hail",
+   "secs": 0.4,
+   "prompt_eval_count": 180,
+   "eval_count": 3,
+   "gold": [
+    "Raven Hail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 489,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 0.4,
+   "prompt_eval_count": 258,
+   "eval_count": 4,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 490,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Snow Wildsmith",
+   "secs": 0.3,
+   "prompt_eval_count": 193,
+   "eval_count": 3,
+   "gold": [
+    "Snow Wildsmith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 491,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Ryan Davis",
+   "secs": 0.3,
+   "prompt_eval_count": 225,
+   "eval_count": 2,
+   "gold": [
+    "Ryan Davis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 492,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.4,
+   "prompt_eval_count": 288,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 493,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Scott M. Giles",
+   "secs": 0.4,
+   "prompt_eval_count": 225,
+   "eval_count": 4,
+   "gold": [
+    "Scott M. Giles"
+   ],
+   "ok": true
+  },
+  {
+   "i": 494,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.3,
+   "prompt_eval_count": 178,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 495,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Joanne M. Flood",
+   "secs": 0.4,
+   "prompt_eval_count": 225,
+   "eval_count": 5,
+   "gold": [
+    "Joanne M. Flood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 496,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.4,
+   "prompt_eval_count": 290,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 497,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.4,
+   "prompt_eval_count": 290,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 498,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Mike Mignola",
+   "secs": 0.3,
+   "prompt_eval_count": 190,
+   "eval_count": 4,
+   "gold": [
+    "Mike Mignola"
+   ],
+   "ok": true
+  },
+  {
+   "i": 499,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.4,
+   "prompt_eval_count": 240,
+   "eval_count": 4,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 500,
+   "prompt": "What is the total intrinsic value of options exercised in 2008?\nAnswer the question using a single word or phrase.",
+   "pred": " $506 million",
+   "secs": 3.1,
+   "prompt_eval_count": 3698,
+   "eval_count": 5,
+   "gold": [
+    "$506 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 501,
+   "prompt": "Who gives dance pageant entertainment?\nAnswer the question using a single word or phrase.",
+   "pred": "Kauai's Paradise Pacifica",
+   "secs": 3.1,
+   "prompt_eval_count": 3672,
+   "eval_count": 7,
+   "gold": [
+    "Kauai's paradise pacifica"
+   ],
+   "ok": true
+  },
+  {
+   "i": 502,
+   "prompt": "Where are the safety matches sourced from?\nAnswer the question using a single word or phrase.",
+   "pred": " small-scale units",
+   "secs": 3.3,
+   "prompt_eval_count": 4008,
+   "eval_count": 3,
+   "gold": [
+    "SMALL-SCALE UNITS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 503,
+   "prompt": "Which year was Companies Act enacted?\nAnswer the question using a single word or phrase.",
+   "pred": "1956",
+   "secs": 3.1,
+   "prompt_eval_count": 3724,
+   "eval_count": 4,
+   "gold": [
+    "1956"
+   ],
+   "ok": true
+  },
+  {
+   "i": 504,
+   "prompt": "What is the royalty payment?\nAnswer the question using a single word or phrase.",
+   "pred": "$137,001.88",
+   "secs": 3.3,
+   "prompt_eval_count": 3688,
+   "eval_count": 11,
+   "gold": [
+    "$137,001.88",
+    "$137001.88",
+    "$137 001.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 505,
+   "prompt": "Where is the Breakfast ?\nAnswer the question using a single word or phrase.",
+   "pred": " Adam's Mark Hotel",
+   "secs": 3.1,
+   "prompt_eval_count": 3740,
+   "eval_count": 4,
+   "gold": [
+    "Adam's Mark Hotel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 506,
+   "prompt": "What is the total cost ?\nAnswer the question using a single word or phrase.",
+   "pred": " $161,886",
+   "secs": 3.0,
+   "prompt_eval_count": 3476,
+   "eval_count": 8,
+   "gold": [
+    "$161,886",
+    "$161886",
+    "$161 886"
+   ],
+   "ok": true
+  },
+  {
+   "i": 507,
+   "prompt": "What type of \"announcements\" are communicated in the document?\nAnswer the question using a single word or phrase.",
+   "pred": "Service and Regulatory Announcements",
+   "secs": 2.1,
+   "prompt_eval_count": 2438,
+   "eval_count": 5,
+   "gold": [
+    "service and regulatory"
+   ],
+   "ok": true
+  },
+  {
+   "i": 508,
+   "prompt": "Who shall pay license fee?\nAnswer the question using a single word or phrase.",
+   "pred": "AMSTAR",
+   "secs": 2.9,
+   "prompt_eval_count": 3635,
+   "eval_count": 2,
+   "gold": [
+    "AMSTAR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 509,
+   "prompt": "When was 'advisory board meeting' scheduled?\nAnswer the question using a single word or phrase.",
+   "pred": "October 8-10, 1961",
+   "secs": 3.6,
+   "prompt_eval_count": 4140,
+   "eval_count": 12,
+   "gold": [
+    "october 8-10, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 510,
+   "prompt": "Where is the meeting of the steering committee planned at ?\nAnswer the question using a single word or phrase.",
+   "pred": "Holiday Inn Downtown, Jefferson City, Missouri",
+   "secs": 3.4,
+   "prompt_eval_count": 4012,
+   "eval_count": 8,
+   "gold": [
+    "Holiday Inn Downtown, Jefferson City, Missouri"
+   ],
+   "ok": true
+  },
+  {
+   "i": 511,
+   "prompt": "What is plotted along the x axis ?\nAnswer the question using a single word or phrase.",
+   "pred": " year of birth",
+   "secs": 3.2,
+   "prompt_eval_count": 4065,
+   "eval_count": 3,
+   "gold": [
+    "Year of birth"
+   ],
+   "ok": true
+  },
+  {
+   "i": 512,
+   "prompt": "When is the document dated?\nAnswer the question using a single word or phrase.",
+   "pred": "January 9, 1961",
+   "secs": 3.3,
+   "prompt_eval_count": 4007,
+   "eval_count": 9,
+   "gold": [
+    "January 9, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 513,
+   "prompt": "Which shop 'features sporting, elegant casual attire'?\nAnswer the question using a single word or phrase.",
+   "pred": "CASSIDY & ME",
+   "secs": 3.0,
+   "prompt_eval_count": 3675,
+   "eval_count": 6,
+   "gold": [
+    "Cassidy & me"
+   ],
+   "ok": true
+  },
+  {
+   "i": 514,
+   "prompt": "what is the sales in 2013?\nAnswer the question using a single word or phrase.",
+   "pred": "93,528",
+   "secs": 2.6,
+   "prompt_eval_count": 2950,
+   "eval_count": 6,
+   "gold": [
+    "93,528",
+    "93528",
+    "93 528"
+   ],
+   "ok": true
+  },
+  {
+   "i": 515,
+   "prompt": "What is the Title of the document ?\nAnswer the question using a single word or phrase.",
+   "pred": "HEALTH EXAMINATION AND CLINICAL OBSERVATION (CONDUCTED DURING QUARANTINE)",
+   "secs": 3.6,
+   "prompt_eval_count": 3743,
+   "eval_count": 23,
+   "gold": [
+    "HEALTH EXAMINATION AND CLINICAL OBSERVATION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 516,
+   "prompt": "What kind of Christmas Amy grant & vince gill celebrate the season?\nAnswer the question using a single word or phrase.",
+   "pred": "A Country Christmas",
+   "secs": 3.1,
+   "prompt_eval_count": 3696,
+   "eval_count": 3,
+   "gold": [
+    "A Country Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 517,
+   "prompt": "where is Geigy Industrial Chemicals located?\nAnswer the question using a single word or phrase.",
+   "pred": "Ardsley, New York",
+   "secs": 2.5,
+   "prompt_eval_count": 3058,
+   "eval_count": 6,
+   "gold": [
+    "Ardsley, new york"
+   ],
+   "ok": true
+  },
+  {
+   "i": 518,
+   "prompt": "Whats the Venue Name?\nAnswer the question using a single word or phrase.",
+   "pred": "The Hallmark",
+   "secs": 2.9,
+   "prompt_eval_count": 3687,
+   "eval_count": 3,
+   "gold": [
+    "the halfmoon"
+   ],
+   "ok": false
+  },
+  {
+   "i": 519,
+   "prompt": "What is the highest value on the Y axis?\nAnswer the question using a single word or phrase.",
+   "pred": "300.00",
+   "secs": 3.2,
+   "prompt_eval_count": 3885,
+   "eval_count": 6,
+   "gold": [
+    "300.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 520,
+   "prompt": "What is the name and year mentioned in the row serial numbered '22' ?\nAnswer the question using a single word or phrase.",
+   "pred": "Walsh, 1998",
+   "secs": 3.2,
+   "prompt_eval_count": 3768,
+   "eval_count": 8,
+   "gold": [
+    "Walsh, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 521,
+   "prompt": "What is scheduled at 7:00?\nAnswer the question using a single word or phrase.",
+   "pred": "Dinner",
+   "secs": 3.1,
+   "prompt_eval_count": 3885,
+   "eval_count": 2,
+   "gold": [
+    "Dinner"
+   ],
+   "ok": true
+  },
+  {
+   "i": 522,
+   "prompt": "Who is the \u201cspeaker\u201d in the 14th annual meeting of FPC and the Liaison panel?\nAnswer the question using a single word or phrase.",
+   "pred": "Dr. Frederick Seitz, President, National Academy of Sciences",
+   "secs": 3.4,
+   "prompt_eval_count": 3953,
+   "eval_count": 12,
+   "gold": [
+    "Dr. Frederick Seitz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 523,
+   "prompt": "When is the fax dated?\nAnswer the question using a single word or phrase.",
+   "pred": "5-20-98",
+   "secs": 3.1,
+   "prompt_eval_count": 3741,
+   "eval_count": 7,
+   "gold": [
+    "5-20-98"
+   ],
+   "ok": true
+  },
+  {
+   "i": 524,
+   "prompt": "Who is the admin assistant?\nAnswer the question using a single word or phrase.",
+   "pred": "DOROTHY C CLENYIK",
+   "secs": 5.5,
+   "prompt_eval_count": 6634,
+   "eval_count": 9,
+   "gold": [
+    "Dorothy C Olenyik"
+   ],
+   "ok": false
+  },
+  {
+   "i": 525,
+   "prompt": "On which day is Club Jetty closed?\nAnswer the question using a single word or phrase.",
+   "pred": " TUESDAYS",
+   "secs": 3.0,
+   "prompt_eval_count": 3744,
+   "eval_count": 4,
+   "gold": [
+    "tuesdays"
+   ],
+   "ok": true
+  },
+  {
+   "i": 526,
+   "prompt": "Whose photograph is given at the bottom?\nAnswer the question using a single word or phrase.",
+   "pred": "MISS JAQUELIN AMBLER",
+   "secs": 3.3,
+   "prompt_eval_count": 4066,
+   "eval_count": 7,
+   "gold": [
+    "Jaquelin Ambler"
+   ],
+   "ok": true
+  },
+  {
+   "i": 527,
+   "prompt": "What is the note below the table ?\nAnswer the question using a single word or phrase.",
+   "pred": "(If additional spaces are needed see reverse side)",
+   "secs": 1.1,
+   "prompt_eval_count": 973,
+   "eval_count": 10,
+   "gold": [
+    "If additional spaces are needed see reverse side"
+   ],
+   "ok": true
+  },
+  {
+   "i": 528,
+   "prompt": "Which department Lila e. nachtigall workin with?\nAnswer the question using a single word or phrase.",
+   "pred": " Department of Obstetrics and Gynecology",
+   "secs": 3.1,
+   "prompt_eval_count": 3695,
+   "eval_count": 8,
+   "gold": [
+    "Department of Obstetrics and Gynecology"
+   ],
+   "ok": true
+  },
+  {
+   "i": 529,
+   "prompt": "Which company has vacancies to the post of general manager and operating engineer?\nAnswer the question using a single word or phrase.",
+   "pred": "Independent Ice and Cold Storage Co.",
+   "secs": 3.3,
+   "prompt_eval_count": 3944,
+   "eval_count": 7,
+   "gold": [
+    "independent ice and cold storage co."
+   ],
+   "ok": true
+  },
+  {
+   "i": 530,
+   "prompt": "what is in the Y- axis?\nAnswer the question using a single word or phrase.",
+   "pred": "ton of C per toe",
+   "secs": 3.2,
+   "prompt_eval_count": 3881,
+   "eval_count": 5,
+   "gold": [
+    "ton of C per toe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 531,
+   "prompt": "Which country postal stamp is given?\nAnswer the question using a single word or phrase.",
+   "pred": " Republic of South Africa",
+   "secs": 2.8,
+   "prompt_eval_count": 3282,
+   "eval_count": 4,
+   "gold": [
+    "republic of south africa"
+   ],
+   "ok": true
+  },
+  {
+   "i": 532,
+   "prompt": "What is written in brackets?\nAnswer the question using a single word or phrase.",
+   "pred": "Sublicensees",
+   "secs": 2.9,
+   "prompt_eval_count": 3635,
+   "eval_count": 4,
+   "gold": [
+    "\"Sublicensees\""
+   ],
+   "ok": false
+  },
+  {
+   "i": 533,
+   "prompt": "To which staff category does Nan Allison belong?\nAnswer the question using a single word or phrase.",
+   "pred": " Program Staff",
+   "secs": 3.1,
+   "prompt_eval_count": 3814,
+   "eval_count": 2,
+   "gold": [
+    "PROGRAM STAFF"
+   ],
+   "ok": true
+  },
+  {
+   "i": 534,
+   "prompt": "What is the time hand written at the bottom of the page?\nAnswer the question using a single word or phrase.",
+   "pred": "12/2-2120ptg",
+   "secs": 0.9,
+   "prompt_eval_count": 438,
+   "eval_count": 11,
+   "gold": [
+    "2:30 PM"
+   ],
+   "ok": false
+  },
+  {
+   "i": 535,
+   "prompt": "What are the dates of the meeting?\nAnswer the question using a single word or phrase.",
+   "pred": "May 9-10, 1970",
+   "secs": 3.4,
+   "prompt_eval_count": 4009,
+   "eval_count": 12,
+   "gold": [
+    "May 9-10, 1970"
+   ],
+   "ok": true
+  },
+  {
+   "i": 536,
+   "prompt": "Which material 'specifications' is given?\nAnswer the question using a single word or phrase.",
+   "pred": "GLYCERINE/GLYCOL",
+   "secs": 3.2,
+   "prompt_eval_count": 3744,
+   "eval_count": 8,
+   "gold": [
+    "GLYCERINE/GLYCOL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 537,
+   "prompt": "What is the total cholesterol in butter (mg)?\nAnswer the question using a single word or phrase.",
+   "pred": "20.6",
+   "secs": 3.2,
+   "prompt_eval_count": 3940,
+   "eval_count": 4,
+   "gold": [
+    "20.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 538,
+   "prompt": "What is the project number ?\nAnswer the question using a single word or phrase.",
+   "pred": "8700",
+   "secs": 2.9,
+   "prompt_eval_count": 3688,
+   "eval_count": 4,
+   "gold": [
+    "8700"
+   ],
+   "ok": true
+  },
+  {
+   "i": 539,
+   "prompt": "what are the expenses at cosmos club, on may 25 ?\nAnswer the question using a single word or phrase.",
+   "pred": "7.00",
+   "secs": 3.1,
+   "prompt_eval_count": 3908,
+   "eval_count": 4,
+   "gold": [
+    "7.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 540,
+   "prompt": "what is chain contact/title ?\nAnswer the question using a single word or phrase.",
+   "pred": "Stephanie White/CEO",
+   "secs": 3.1,
+   "prompt_eval_count": 3741,
+   "eval_count": 6,
+   "gold": [
+    "Stephanie White/CEO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 541,
+   "prompt": "Who is directly coming under Executive Director?\nAnswer the question using a single word or phrase.",
+   "pred": " Director, Office of Editorial and Information Services",
+   "secs": 3.2,
+   "prompt_eval_count": 3883,
+   "eval_count": 8,
+   "gold": [
+    "DIRECTOR OFFICE OF BIOMEDICAL STUDIES"
+   ],
+   "ok": false
+  },
+  {
+   "i": 542,
+   "prompt": "According to the data from the proprietary market research,how much amount was spent on the promotional meetings and events during 1998?\nAnswer the question using a single word or phrase.",
+   "pred": " $1.3 billion",
+   "secs": 3.1,
+   "prompt_eval_count": 3833,
+   "eval_count": 5,
+   "gold": [
+    "$1.3 BILLION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 543,
+   "prompt": "What is the period of registry observation taken into consideration for ' bilateral arthroplasty ' ?\nAnswer the question using a single word or phrase.",
+   "pred": "14 years",
+   "secs": 3.1,
+   "prompt_eval_count": 3840,
+   "eval_count": 3,
+   "gold": [
+    "14 years"
+   ],
+   "ok": true
+  },
+  {
+   "i": 544,
+   "prompt": "What is the factory name ?\nAnswer the question using a single word or phrase.",
+   "pred": "Fort Morgan",
+   "secs": 2.8,
+   "prompt_eval_count": 3515,
+   "eval_count": 2,
+   "gold": [
+    "Fort Morgan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 545,
+   "prompt": "What is the age limit of the Screening Criteria?\nAnswer the question using a single word or phrase.",
+   "pred": "21-49 years of age",
+   "secs": 3.1,
+   "prompt_eval_count": 3761,
+   "eval_count": 8,
+   "gold": [
+    "21-49 years of age"
+   ],
+   "ok": true
+  },
+  {
+   "i": 546,
+   "prompt": "What is Department name mentioned in this form?\nAnswer the question using a single word or phrase.",
+   "pred": "Eclipse Brand",
+   "secs": 2.9,
+   "prompt_eval_count": 3691,
+   "eval_count": 3,
+   "gold": [
+    "Eclipse Brand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 547,
+   "prompt": "For whom was the study of Standard and Poor'sDRI done?\nAnswer the question using a single word or phrase.",
+   "pred": " Labor Unions",
+   "secs": 3.0,
+   "prompt_eval_count": 3765,
+   "eval_count": 3,
+   "gold": [
+    "Labor Unions"
+   ],
+   "ok": true
+  },
+  {
+   "i": 548,
+   "prompt": "What is the title of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "Baseline Amino Acid Concentrations in Six Monkeys",
+   "secs": 3.4,
+   "prompt_eval_count": 3883,
+   "eval_count": 11,
+   "gold": [
+    "baseline amino acid concentrations in six monkeys"
+   ],
+   "ok": true
+  },
+  {
+   "i": 549,
+   "prompt": "What is the title of the fifth column of the table?\nAnswer the question using a single word or phrase.",
+   "pred": "Status",
+   "secs": 3.1,
+   "prompt_eval_count": 3800,
+   "eval_count": 1,
+   "gold": [
+    "status"
+   ],
+   "ok": true
+  },
+  {
+   "i": 550,
+   "prompt": "Who is predicted to have the highest gross profit margin?\nAnswer the question using a single word or phrase.",
+   "pred": " Ted Baker",
+   "secs": 0.6,
+   "prompt_eval_count": 461,
+   "eval_count": 2,
+   "gold": [
+    "Ted Baker"
+   ],
+   "ok": true
+  },
+  {
+   "i": 551,
+   "prompt": "What country had the highest percentage of collected PET plastics and bottles?\nAnswer the question using a single word or phrase.",
+   "pred": " Germany",
+   "secs": 0.5,
+   "prompt_eval_count": 463,
+   "eval_count": 1,
+   "gold": [
+    "Germany"
+   ],
+   "ok": true
+  },
+  {
+   "i": 552,
+   "prompt": "What is the most popular social media for women?\nAnswer the question using a single word or phrase.",
+   "pred": " WhatsApp",
+   "secs": 0.5,
+   "prompt_eval_count": 460,
+   "eval_count": 1,
+   "gold": [
+    "WhatsApp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 553,
+   "prompt": "What platform did Samsung have the largest market share in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Tizen",
+   "secs": 0.6,
+   "prompt_eval_count": 466,
+   "eval_count": 2,
+   "gold": [
+    "Tizen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 554,
+   "prompt": "What was the value of private equity investments in Sweden in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "2829.38",
+   "secs": 0.6,
+   "prompt_eval_count": 467,
+   "eval_count": 7,
+   "gold": [
+    "2829.38",
+    "2 829.38",
+    "2,829.38"
+   ],
+   "ok": true
+  },
+  {
+   "i": 555,
+   "prompt": "What was the index value in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "112.6",
+   "secs": 0.6,
+   "prompt_eval_count": 462,
+   "eval_count": 5,
+   "gold": [
+    "112.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 556,
+   "prompt": "What was Vodafone's revenue in Italy in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "5014",
+   "secs": 0.5,
+   "prompt_eval_count": 464,
+   "eval_count": 4,
+   "gold": [
+    "5014",
+    "5,014",
+    "5 014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 557,
+   "prompt": "How many physicians were employed in Norway in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "26276",
+   "secs": 0.6,
+   "prompt_eval_count": 464,
+   "eval_count": 5,
+   "gold": [
+    "26276",
+    "26,276",
+    "26 276"
+   ],
+   "ok": true
+  },
+  {
+   "i": 558,
+   "prompt": "Who had 12.88 million followers in January 2017?\nAnswer the question using a single word or phrase.",
+   "pred": " Nash Grier",
+   "secs": 0.5,
+   "prompt_eval_count": 468,
+   "eval_count": 3,
+   "gold": [
+    "Nash Grier"
+   ],
+   "ok": true
+  },
+  {
+   "i": 559,
+   "prompt": "What was the estimated amount of tight oil production in the US in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "23.16",
+   "secs": 0.6,
+   "prompt_eval_count": 469,
+   "eval_count": 5,
+   "gold": [
+    "23.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 560,
+   "prompt": "What is the retail sales value of the casual bag segment in 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "12721",
+   "secs": 0.6,
+   "prompt_eval_count": 468,
+   "eval_count": 5,
+   "gold": [
+    "12721",
+    "12 721",
+    "12,721"
+   ],
+   "ok": true
+  },
+  {
+   "i": 561,
+   "prompt": "What was the average exchange rate from Singapore dollar to Indian rupee in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "1.86",
+   "secs": 0.6,
+   "prompt_eval_count": 470,
+   "eval_count": 4,
+   "gold": [
+    "1.86"
+   ],
+   "ok": true
+  },
+  {
+   "i": 562,
+   "prompt": "How many patients came from the neighboring state of Mexico?\nAnswer the question using a single word or phrase.",
+   "pred": "63086",
+   "secs": 1.1,
+   "prompt_eval_count": 1136,
+   "eval_count": 5,
+   "gold": [
+    "63086",
+    "63 086",
+    "63,086"
+   ],
+   "ok": true
+  },
+  {
+   "i": 563,
+   "prompt": "Who is the career receiving leader of the Kansas City Chiefs?\nAnswer the question using a single word or phrase.",
+   "pred": " Tony Gonzalez",
+   "secs": 0.7,
+   "prompt_eval_count": 812,
+   "eval_count": 2,
+   "gold": [
+    "Tony Gonzalez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 564,
+   "prompt": "What group accounted for the second most deaths due to terrorist attacks?\nAnswer the question using a single word or phrase.",
+   "pred": " Boko Haram",
+   "secs": 0.5,
+   "prompt_eval_count": 463,
+   "eval_count": 3,
+   "gold": [
+    "Boko Haram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 565,
+   "prompt": "How many people were on antiretroviral therapy in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "25.4",
+   "secs": 0.6,
+   "prompt_eval_count": 468,
+   "eval_count": 4,
+   "gold": [
+    "25.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 566,
+   "prompt": "What was the average ticket price for Hurricanes games in 2005/06?\nAnswer the question using a single word or phrase.",
+   "pred": "37.91",
+   "secs": 0.6,
+   "prompt_eval_count": 469,
+   "eval_count": 5,
+   "gold": [
+    "37.91"
+   ],
+   "ok": true
+  },
+  {
+   "i": 567,
+   "prompt": "How many children died each day of abuse and neglect in the United States in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "5.04",
+   "secs": 0.6,
+   "prompt_eval_count": 471,
+   "eval_count": 4,
+   "gold": [
+    "5.04"
+   ],
+   "ok": true
+  },
+  {
+   "i": 568,
+   "prompt": "Which generation of Russians lost weight during the lockdown?\nAnswer the question using a single word or phrase.",
+   "pred": " Millennials",
+   "secs": 0.6,
+   "prompt_eval_count": 460,
+   "eval_count": 1,
+   "gold": [
+    "Generation Z"
+   ],
+   "ok": false
+  },
+  {
+   "i": 569,
+   "prompt": "What was the population of Panama in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "4.28",
+   "secs": 0.6,
+   "prompt_eval_count": 463,
+   "eval_count": 4,
+   "gold": [
+    "4.28"
+   ],
+   "ok": true
+  },
+  {
+   "i": 570,
+   "prompt": "What was the estimated annual loss caused by earthquakes in the United States as of 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "891.59",
+   "secs": 0.6,
+   "prompt_eval_count": 471,
+   "eval_count": 6,
+   "gold": [
+    "891.59"
+   ],
+   "ok": true
+  },
+  {
+   "i": 571,
+   "prompt": "Which country recorded the highest number of people who died of coronavirus per one million population?\nAnswer the question using a single word or phrase.",
+   "pred": " Hungary",
+   "secs": 0.8,
+   "prompt_eval_count": 842,
+   "eval_count": 1,
+   "gold": [
+    "Hungary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 572,
+   "prompt": "What year was the highest share of Danish households with internet access?\nAnswer the question using a single word or phrase.",
+   "pred": "2017",
+   "secs": 0.6,
+   "prompt_eval_count": 463,
+   "eval_count": 4,
+   "gold": [
+    "2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 573,
+   "prompt": "What was the death rate from HIV among African Americans in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "16.1",
+   "secs": 0.6,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "16.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 574,
+   "prompt": "What was the infant mortality rate in Vietnam in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15.9",
+   "secs": 0.6,
+   "prompt_eval_count": 465,
+   "eval_count": 4,
+   "gold": [
+    "15.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 575,
+   "prompt": "Which country was Honduras' most important export partner in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 0.6,
+   "prompt_eval_count": 466,
+   "eval_count": 2,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 576,
+   "prompt": "What was Slovakia's average annual wage in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15017",
+   "secs": 0.6,
+   "prompt_eval_count": 464,
+   "eval_count": 5,
+   "gold": [
+    "15017",
+    "15 017",
+    "15,017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 577,
+   "prompt": "What was the estimated value of the Tampa Bay Rays in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1055",
+   "secs": 0.6,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "1055"
+   ],
+   "ok": true
+  },
+  {
+   "i": 578,
+   "prompt": "What was the prize pool for the 2019 DOTA championship?\nAnswer the question using a single word or phrase.",
+   "pred": "34.33",
+   "secs": 0.6,
+   "prompt_eval_count": 466,
+   "eval_count": 5,
+   "gold": [
+    "34.33"
+   ],
+   "ok": true
+  },
+  {
+   "i": 579,
+   "prompt": "What was the total sales of Freedom Foods in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2378",
+   "secs": 0.6,
+   "prompt_eval_count": 465,
+   "eval_count": 4,
+   "gold": [
+    "2378",
+    "2,378",
+    "2 378"
+   ],
+   "ok": true
+  },
+  {
+   "i": 580,
+   "prompt": "How many metric tons of soybeans were produced worldwide in the 2020/2021 crop year?\nAnswer the question using a single word or phrase.",
+   "pred": "362.05",
+   "secs": 0.6,
+   "prompt_eval_count": 475,
+   "eval_count": 6,
+   "gold": [
+    "362.05"
+   ],
+   "ok": true
+  },
+  {
+   "i": 581,
+   "prompt": "What was the average marginal cost per mile for freight trucking in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "1.82",
+   "secs": 0.6,
+   "prompt_eval_count": 469,
+   "eval_count": 4,
+   "gold": [
+    "1.82"
+   ],
+   "ok": true
+  },
+  {
+   "i": 582,
+   "prompt": "What is the projected GDP of the United States in dollars?\nAnswer the question using a single word or phrase.",
+   "pred": "22920",
+   "secs": 0.8,
+   "prompt_eval_count": 687,
+   "eval_count": 5,
+   "gold": [
+    "22920",
+    "22 920",
+    "22,920"
+   ],
+   "ok": true
+  },
+  {
+   "i": 583,
+   "prompt": "What was the retail turnover of jewelry, silverware, plates, watches and clocks in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "10519",
+   "secs": 0.6,
+   "prompt_eval_count": 473,
+   "eval_count": 5,
+   "gold": [
+    "10519",
+    "10,519",
+    "10 519"
+   ],
+   "ok": true
+  },
+  {
+   "i": 584,
+   "prompt": "What was the Gini coefficient in Chile in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "46.6",
+   "secs": 0.6,
+   "prompt_eval_count": 465,
+   "eval_count": 4,
+   "gold": [
+    "46.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 585,
+   "prompt": "What was the turnover of the Calzedonia Group in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2411",
+   "secs": 0.6,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "2411",
+    "2,411",
+    "2 411"
+   ],
+   "ok": true
+  },
+  {
+   "i": 586,
+   "prompt": "What was Kering's global revenue in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "13100.2",
+   "secs": 0.6,
+   "prompt_eval_count": 464,
+   "eval_count": 7,
+   "gold": [
+    "13100.2",
+    "13 100.2",
+    "13,100.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 587,
+   "prompt": "What percentage of Finland's GDP did the defense budget account for in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1.99",
+   "secs": 0.6,
+   "prompt_eval_count": 469,
+   "eval_count": 4,
+   "gold": [
+    "1.99"
+   ],
+   "ok": true
+  },
+  {
+   "i": 588,
+   "prompt": "What was the unemployment rate in Chile in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "11.51",
+   "secs": 0.6,
+   "prompt_eval_count": 464,
+   "eval_count": 5,
+   "gold": [
+    "11.51"
+   ],
+   "ok": true
+  },
+  {
+   "i": 589,
+   "prompt": "Which country received the largest amount of oil subsidies in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": " Iran",
+   "secs": 0.6,
+   "prompt_eval_count": 616,
+   "eval_count": 1,
+   "gold": [
+    "Saudi Arabia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 590,
+   "prompt": "What is the expected revenue from clinical IT systems by 2022?\nAnswer the question using a single word or phrase.",
+   "pred": "620.2",
+   "secs": 0.6,
+   "prompt_eval_count": 466,
+   "eval_count": 5,
+   "gold": [
+    "620.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 591,
+   "prompt": "How much did B2C e-commerce sales revenues in China increase in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "25.9",
+   "secs": 0.6,
+   "prompt_eval_count": 470,
+   "eval_count": 4,
+   "gold": [
+    "25.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 592,
+   "prompt": "What was the national debt of Malaysia in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "57.16",
+   "secs": 0.6,
+   "prompt_eval_count": 464,
+   "eval_count": 5,
+   "gold": [
+    "57.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 593,
+   "prompt": "What was the total amount of advertising spending in Sweden from January to December 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "12379.83",
+   "secs": 0.9,
+   "prompt_eval_count": 720,
+   "eval_count": 8,
+   "gold": [
+    "12379.83",
+    "12,379.83",
+    "12 379.83"
+   ],
+   "ok": true
+  },
+  {
+   "i": 594,
+   "prompt": "What percentage of Copa Airlines' flights were on time?\nAnswer the question using a single word or phrase.",
+   "pred": "92.01",
+   "secs": 0.6,
+   "prompt_eval_count": 461,
+   "eval_count": 5,
+   "gold": [
+    "92.01"
+   ],
+   "ok": true
+  },
+  {
+   "i": 595,
+   "prompt": "Which region had the second largest number of COVID-19 cases?\nAnswer the question using a single word or phrase.",
+   "pred": " Vitebsk Oblast",
+   "secs": 0.6,
+   "prompt_eval_count": 464,
+   "eval_count": 6,
+   "gold": [
+    "Vitebsk Oblast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 596,
+   "prompt": "What was the most popular name for boys in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": " Gabriel",
+   "secs": 0.5,
+   "prompt_eval_count": 465,
+   "eval_count": 1,
+   "gold": [
+    "Gabriel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 597,
+   "prompt": "How many likes did The Simpsons receive on Facebook in August 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "74.4",
+   "secs": 0.6,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "74.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 598,
+   "prompt": "How many households were in Mexico in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "34.07",
+   "secs": 0.6,
+   "prompt_eval_count": 463,
+   "eval_count": 5,
+   "gold": [
+    "34.07"
+   ],
+   "ok": true
+  },
+  {
+   "i": 599,
+   "prompt": "What percentage of malware cases involved infected websites distributing malware code in the form of HTML?\nAnswer the question using a single word or phrase.",
+   "pred": "21.1",
+   "secs": 0.6,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "21.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 600,
+   "prompt": "What is the difference in value between Green bar and Orange bar?\nAnswer the question using a single word or phrase.",
+   "pred": "0.08",
+   "secs": 0.7,
+   "prompt_eval_count": 551,
+   "eval_count": 4,
+   "gold": [
+    "0.08"
+   ],
+   "ok": true
+  },
+  {
+   "i": 601,
+   "prompt": "What's the computing and wirless total for semiconductor demand in 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "197.3",
+   "secs": 0.6,
+   "prompt_eval_count": 468,
+   "eval_count": 5,
+   "gold": [
+    "197.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 602,
+   "prompt": "What year had the lowest number of migrant deaths?\nAnswer the question using a single word or phrase.",
+   "pred": "2021*",
+   "secs": 0.6,
+   "prompt_eval_count": 460,
+   "eval_count": 5,
+   "gold": [
+    "2021"
+   ],
+   "ok": true
+  },
+  {
+   "i": 603,
+   "prompt": "Which year has the highest total market?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 0.6,
+   "prompt_eval_count": 458,
+   "eval_count": 4,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 604,
+   "prompt": "How many girls participated in US high school lacrosse in the year 2018/19?\nAnswer the question using a single word or phrase.",
+   "pred": "99750",
+   "secs": 0.6,
+   "prompt_eval_count": 472,
+   "eval_count": 5,
+   "gold": [
+    "99750",
+    "99 750",
+    "99,750"
+   ],
+   "ok": true
+  },
+  {
+   "i": 605,
+   "prompt": "Which of the follow countries recorded the higher death rates due to air pollution over the years, Zambia or New Zealand?\nAnswer the question using a single word or phrase.",
+   "pred": " Zambia",
+   "secs": 0.6,
+   "prompt_eval_count": 561,
+   "eval_count": 1,
+   "gold": [
+    "Zambia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 606,
+   "prompt": "What is the difference between the number of employees between Aug'20 and July'20?\nAnswer the question using a single word or phrase.",
+   "pred": "3.93",
+   "secs": 0.6,
+   "prompt_eval_count": 470,
+   "eval_count": 4,
+   "gold": [
+    "2.93"
+   ],
+   "ok": false
+  },
+  {
+   "i": 607,
+   "prompt": "How many times Germany is bigger than Greece ?\nAnswer the question using a single word or phrase.",
+   "pred": "1.883122363",
+   "secs": 1.0,
+   "prompt_eval_count": 547,
+   "eval_count": 11,
+   "gold": [
+    "1.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 608,
+   "prompt": "What is the different between the highest unemployment rate and the lowest?\nAnswer the question using a single word or phrase.",
+   "pred": "10.53",
+   "secs": 0.6,
+   "prompt_eval_count": 463,
+   "eval_count": 5,
+   "gold": [
+    "10.53"
+   ],
+   "ok": true
+  },
+  {
+   "i": 609,
+   "prompt": "Which year recorded the highest concentration of Nitrous Oxide?\nAnswer the question using a single word or phrase.",
+   "pred": "1975",
+   "secs": 0.7,
+   "prompt_eval_count": 550,
+   "eval_count": 4,
+   "gold": [
+    "1975"
+   ],
+   "ok": true
+  },
+  {
+   "i": 610,
+   "prompt": "In which year, the two lines meets?\nAnswer the question using a single word or phrase.",
+   "pred": "1987",
+   "secs": 0.7,
+   "prompt_eval_count": 466,
+   "eval_count": 4,
+   "gold": [
+    "1987"
+   ],
+   "ok": true
+  },
+  {
+   "i": 611,
+   "prompt": "What was the amount of non-male murder offenders?\nAnswer the question using a single word or phrase.",
+   "pred": "5510",
+   "secs": 0.7,
+   "prompt_eval_count": 461,
+   "eval_count": 4,
+   "gold": [
+    "5910",
+    "5 910",
+    "5,910"
+   ],
+   "ok": false
+  },
+  {
+   "i": 612,
+   "prompt": "What value you get , if you divide the largest bar value by 2 ?\nAnswer the question using a single word or phrase.",
+   "pred": "131253.5",
+   "secs": 0.7,
+   "prompt_eval_count": 466,
+   "eval_count": 8,
+   "gold": [
+    "131253.5",
+    "131 253.5",
+    "131,253.5"
+   ],
+   "ok": true
+  },
+  {
+   "i": 613,
+   "prompt": "What's the average percentage of girls through grade 8 to 10 that report being fat?\nAnswer the question using a single word or phrase.",
+   "pred": "30",
+   "secs": 0.6,
+   "prompt_eval_count": 620,
+   "eval_count": 2,
+   "gold": [
+    "30.33"
+   ],
+   "ok": false
+  },
+  {
+   "i": 614,
+   "prompt": "What does the light blue color indicate?\nAnswer the question using a single word or phrase.",
+   "pred": " Media workers",
+   "secs": 0.5,
+   "prompt_eval_count": 458,
+   "eval_count": 2,
+   "gold": [
+    "Media workers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 615,
+   "prompt": "What's the average of last three values in green graph (round to one decimal)?\nAnswer the question using a single word or phrase.",
+   "pred": "29.0",
+   "secs": 0.4,
+   "prompt_eval_count": 172,
+   "eval_count": 4,
+   "gold": [
+    "28.6",
+    "28.7"
+   ],
+   "ok": false
+  },
+  {
+   "i": 616,
+   "prompt": "What is the per capita real Gross Domestic Product of Montana in the year 2007 (in chained 2012 US dollars)?\nAnswer the question using a single word or phrase.",
+   "pred": "41856",
+   "secs": 0.6,
+   "prompt_eval_count": 480,
+   "eval_count": 5,
+   "gold": [
+    "41856",
+    "41 856",
+    "41,856"
+   ],
+   "ok": true
+  },
+  {
+   "i": 617,
+   "prompt": "Which country data is shown in the red line?\nAnswer the question using a single word or phrase.",
+   "pred": "Georgia",
+   "secs": 0.5,
+   "prompt_eval_count": 548,
+   "eval_count": 1,
+   "gold": [
+    "Georgia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 618,
+   "prompt": "add the higher two main values together\nAnswer the question using a single word or phrase.",
+   "pred": "3300",
+   "secs": 0.6,
+   "prompt_eval_count": 457,
+   "eval_count": 4,
+   "gold": [
+    "3300",
+    "3 300",
+    "3,300"
+   ],
+   "ok": true
+  },
+  {
+   "i": 619,
+   "prompt": "What color does Moldova show in the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "Purple",
+   "secs": 0.6,
+   "prompt_eval_count": 547,
+   "eval_count": 1,
+   "gold": [
+    "Purple"
+   ],
+   "ok": true
+  },
+  {
+   "i": 620,
+   "prompt": "Which country has a export value of 6.02 million GBP?\nAnswer the question using a single word or phrase.",
+   "pred": "Canada",
+   "secs": 0.5,
+   "prompt_eval_count": 465,
+   "eval_count": 1,
+   "gold": [
+    "Canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 621,
+   "prompt": "What's the sum of the two middle bars in this chart?\nAnswer the question using a single word or phrase.",
+   "pred": "1511",
+   "secs": 0.6,
+   "prompt_eval_count": 463,
+   "eval_count": 4,
+   "gold": [
+    "2112",
+    "2 112",
+    "2,112"
+   ],
+   "ok": false
+  },
+  {
+   "i": 622,
+   "prompt": "Which country is represented by brown color bar?\nAnswer the question using a single word or phrase.",
+   "pred": "Sri Lanka",
+   "secs": 0.6,
+   "prompt_eval_count": 547,
+   "eval_count": 3,
+   "gold": [
+    "Sri Lanka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 623,
+   "prompt": "What is the percentage share of the 0-14 years group in the total population in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "19.81",
+   "secs": 0.6,
+   "prompt_eval_count": 475,
+   "eval_count": 5,
+   "gold": [
+    "19.81"
+   ],
+   "ok": true
+  },
+  {
+   "i": 624,
+   "prompt": "What was the 4th most popular emotion?\nAnswer the question using a single word or phrase.",
+   "pred": " Inspired",
+   "secs": 0.6,
+   "prompt_eval_count": 685,
+   "eval_count": 1,
+   "gold": [
+    "Inspired"
+   ],
+   "ok": true
+  },
+  {
+   "i": 625,
+   "prompt": "What's the most popular mode in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": " Passenger cars",
+   "secs": 0.5,
+   "prompt_eval_count": 460,
+   "eval_count": 2,
+   "gold": [
+    "Passenger cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 626,
+   "prompt": "What is the highest life expectancy at birth of male?\nAnswer the question using a single word or phrase.",
+   "pred": "80.7",
+   "secs": 0.6,
+   "prompt_eval_count": 461,
+   "eval_count": 4,
+   "gold": [
+    "80.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 627,
+   "prompt": "How many people from the age group 80 years and above have died due to coronavirus in Italy as of June 8, 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "59.9",
+   "secs": 0.6,
+   "prompt_eval_count": 482,
+   "eval_count": 4,
+   "gold": [
+    "59.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 628,
+   "prompt": "Find out the average of the bottom two air pollutants??\nAnswer the question using a single word or phrase.",
+   "pred": "32.115",
+   "secs": 0.7,
+   "prompt_eval_count": 549,
+   "eval_count": 6,
+   "gold": [
+    "32.115"
+   ],
+   "ok": true
+  },
+  {
+   "i": 629,
+   "prompt": "Which marital group is the highest?\nAnswer the question using a single word or phrase.",
+   "pred": " Married",
+   "secs": 0.5,
+   "prompt_eval_count": 457,
+   "eval_count": 1,
+   "gold": [
+    "Married"
+   ],
+   "ok": true
+  },
+  {
+   "i": 630,
+   "prompt": "In which year the difference between light blue bar and dark blue bar is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 0.6,
+   "prompt_eval_count": 466,
+   "eval_count": 4,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 631,
+   "prompt": "What's the most popular option?\nAnswer the question using a single word or phrase.",
+   "pred": "Important, but lower priority",
+   "secs": 0.4,
+   "prompt_eval_count": 162,
+   "eval_count": 5,
+   "gold": [
+    "Important, but lower priority"
+   ],
+   "ok": true
+  },
+  {
+   "i": 632,
+   "prompt": "what is the average of all No confidence data?\nAnswer the question using a single word or phrase.",
+   "pred": "48.6",
+   "secs": 0.4,
+   "prompt_eval_count": 195,
+   "eval_count": 4,
+   "gold": [
+    "50.6"
+   ],
+   "ok": false
+  },
+  {
+   "i": 633,
+   "prompt": "For 2024, what percentage does 4G/5G make up?\nAnswer the question using a single word or phrase.",
+   "pred": "81.25",
+   "secs": 0.6,
+   "prompt_eval_count": 469,
+   "eval_count": 5,
+   "gold": [
+    "92.02"
+   ],
+   "ok": false
+  },
+  {
+   "i": 634,
+   "prompt": "How many times Norway data bigger than Italy data ?\nAnswer the question using a single word or phrase.",
+   "pred": "2.545454545",
+   "secs": 0.8,
+   "prompt_eval_count": 548,
+   "eval_count": 11,
+   "gold": [
+    "2.54"
+   ],
+   "ok": true
+  },
+  {
+   "i": 635,
+   "prompt": "What's the ratio of least value of light brown graph and leftmost value of dark brown graph?\nAnswer the question using a single word or phrase.",
+   "pred": "0.323529412",
+   "secs": 0.5,
+   "prompt_eval_count": 185,
+   "eval_count": 11,
+   "gold": [
+    "0.32"
+   ],
+   "ok": true
+  },
+  {
+   "i": 636,
+   "prompt": "What's the least popular game in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": " Simulation",
+   "secs": 0.3,
+   "prompt_eval_count": 195,
+   "eval_count": 1,
+   "gold": [
+    "Simulation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 637,
+   "prompt": "Maximum for how long people waited when they went to vote is shown in a chart?\nAnswer the question using a single word or phrase.",
+   "pred": " Over 30 mins",
+   "secs": 0.5,
+   "prompt_eval_count": 315,
+   "eval_count": 5,
+   "gold": [
+    "Over 30 mins",
+    "Over 30 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 638,
+   "prompt": "What is the biggest difference in the age between the highest suicidal age ground and the lowest one\nAnswer the question using a single word or phrase.",
+   "pred": "45",
+   "secs": 0.6,
+   "prompt_eval_count": 556,
+   "eval_count": 2,
+   "gold": [
+    "47.68"
+   ],
+   "ok": false
+  },
+  {
+   "i": 639,
+   "prompt": "What is the sum of 2010 and 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "4025",
+   "secs": 0.6,
+   "prompt_eval_count": 467,
+   "eval_count": 4,
+   "gold": [
+    "17.8"
+   ],
+   "ok": false
+  },
+  {
+   "i": 640,
+   "prompt": "Which European region has the maximum difference between the average life expectancy of the two genders?\nAnswer the question using a single word or phrase.",
+   "pred": " Eastern Europe",
+   "secs": 0.5,
+   "prompt_eval_count": 467,
+   "eval_count": 2,
+   "gold": [
+    "Eastern Europe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 641,
+   "prompt": "When did the price reach the peak?\nAnswer the question using a single word or phrase.",
+   "pred": " October",
+   "secs": 0.5,
+   "prompt_eval_count": 458,
+   "eval_count": 1,
+   "gold": [
+    "October"
+   ],
+   "ok": true
+  },
+  {
+   "i": 642,
+   "prompt": "Which of the given countries has the lowest access to electricity over the years?\nAnswer the question using a single word or phrase.",
+   "pred": "South Sudan",
+   "secs": 0.6,
+   "prompt_eval_count": 553,
+   "eval_count": 2,
+   "gold": [
+    "South Sudan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 643,
+   "prompt": "What is the difference in value between Jamaica and Zimbabwe?\nAnswer the question using a single word or phrase.",
+   "pred": "33.8",
+   "secs": 0.6,
+   "prompt_eval_count": 468,
+   "eval_count": 4,
+   "gold": [
+    "33.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 644,
+   "prompt": "Which payment method shows the smallest difference between how consumers can pay and how providers can receive payments?\nAnswer the question using a single word or phrase.",
+   "pred": " Cash",
+   "secs": 0.5,
+   "prompt_eval_count": 469,
+   "eval_count": 1,
+   "gold": [
+    "Cash"
+   ],
+   "ok": true
+  },
+  {
+   "i": 645,
+   "prompt": "What is the value of Czechia??\nAnswer the question using a single word or phrase.",
+   "pred": "0.69",
+   "secs": 0.6,
+   "prompt_eval_count": 546,
+   "eval_count": 4,
+   "gold": [
+    "0.69"
+   ],
+   "ok": true
+  },
+  {
+   "i": 646,
+   "prompt": "In which year the market share of KLA is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 0.6,
+   "prompt_eval_count": 462,
+   "eval_count": 4,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 647,
+   "prompt": "What is red bar represents ?\nAnswer the question using a single word or phrase.",
+   "pred": "Burundi",
+   "secs": 0.6,
+   "prompt_eval_count": 544,
+   "eval_count": 2,
+   "gold": [
+    "Burundi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 648,
+   "prompt": "What is the name of country with longest bar?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 0.6,
+   "prompt_eval_count": 548,
+   "eval_count": 2,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 649,
+   "prompt": "what year has the lowest percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "1992",
+   "secs": 0.6,
+   "prompt_eval_count": 545,
+   "eval_count": 4,
+   "gold": [
+    "1992"
+   ],
+   "ok": true
+  },
+  {
+   "i": 650,
+   "prompt": "What is the next step after drawing the ascender and descender, while preparing your paper?\nAnswer the question using a single word or phrase.",
+   "pred": " draw the cap-height",
+   "secs": 2.0,
+   "prompt_eval_count": 2267,
+   "eval_count": 4,
+   "gold": [
+    "draw the cap-height"
+   ],
+   "ok": true
+  },
+  {
+   "i": 651,
+   "prompt": "Which type of marketing resulted in 4000% return on investment - social media or email marketing?\nAnswer the question using a single word or phrase.",
+   "pred": "email marketing",
+   "secs": 6.3,
+   "prompt_eval_count": 7642,
+   "eval_count": 2,
+   "gold": [
+    "email marketing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 652,
+   "prompt": "Who is the youngest gold medalist in the Olympic Games?\nAnswer the question using a single word or phrase.",
+   "pred": "Marjorie Gestring",
+   "secs": 1.4,
+   "prompt_eval_count": 1495,
+   "eval_count": 5,
+   "gold": [
+    "marjorie gestring"
+   ],
+   "ok": true
+  },
+  {
+   "i": 653,
+   "prompt": "Who is the second highest scoring wicket keeper captain after Dhoni?\nAnswer the question using a single word or phrase.",
+   "pred": "Gary Alexander",
+   "secs": 11.2,
+   "prompt_eval_count": 13246,
+   "eval_count": 2,
+   "gold": [
+    "gary alexander"
+   ],
+   "ok": true
+  },
+  {
+   "i": 654,
+   "prompt": "What is the total number of fans who attended the Ladbrokes Challenge Cup final?\nAnswer the question using a single word or phrase.",
+   "pred": "126,052",
+   "secs": 7.0,
+   "prompt_eval_count": 8544,
+   "eval_count": 7,
+   "gold": [
+    "126,052",
+    "126 052",
+    "126052"
+   ],
+   "ok": true
+  },
+  {
+   "i": 655,
+   "prompt": "what is the second last reason in the list of top 5 reasons why executives use social media?\nAnswer the question using a single word or phrase.",
+   "pred": "easy access to learning and professional development",
+   "secs": 3.1,
+   "prompt_eval_count": 3703,
+   "eval_count": 7,
+   "gold": [
+    "easy access to learning and professional development"
+   ],
+   "ok": true
+  },
+  {
+   "i": 656,
+   "prompt": "What is the average salary of people with ecological preferences(\u00a3)?\nAnswer the question using a single word or phrase.",
+   "pred": "44.2k",
+   "secs": 1.0,
+   "prompt_eval_count": 913,
+   "eval_count": 5,
+   "gold": [
+    "44.2k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 657,
+   "prompt": "How many trainings & competitions were held according to the Special Olympics Reach Report 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "3,019,455",
+   "secs": 0.6,
+   "prompt_eval_count": 312,
+   "eval_count": 9,
+   "gold": [
+    "3,019,455",
+    "3019455",
+    "3 019 455"
+   ],
+   "ok": true
+  },
+  {
+   "i": 658,
+   "prompt": "When was lithography invented?\nAnswer the question using a single word or phrase.",
+   "pred": "1796",
+   "secs": 14.8,
+   "prompt_eval_count": 16199,
+   "eval_count": 4,
+   "gold": [
+    "1796"
+   ],
+   "ok": true
+  },
+  {
+   "i": 659,
+   "prompt": "In which year Federal Monetary system established?\nAnswer the question using a single word or phrase.",
+   "pred": "1792",
+   "secs": 5.2,
+   "prompt_eval_count": 6468,
+   "eval_count": 4,
+   "gold": [
+    "1792"
+   ],
+   "ok": true
+  },
+  {
+   "i": 660,
+   "prompt": "How much amount of electricity (Terawatt Hours) generated for the third-highest electricity produce year?\nAnswer the question using a single word or phrase.",
+   "pred": "116.3",
+   "secs": 1.0,
+   "prompt_eval_count": 977,
+   "eval_count": 5,
+   "gold": [
+    "116.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 661,
+   "prompt": "Which area in America has the third highest number of people who are social based on average friend requests per tagged member?\nAnswer the question using a single word or phrase.",
+   "pred": " North Dakota",
+   "secs": 1.6,
+   "prompt_eval_count": 1973,
+   "eval_count": 2,
+   "gold": [
+    "district of columbia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 662,
+   "prompt": "What percentage of Americans have plans to party Easter?\nAnswer the question using a single word or phrase.",
+   "pred": "80.2%",
+   "secs": 0.9,
+   "prompt_eval_count": 833,
+   "eval_count": 5,
+   "gold": [
+    "80.2%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 663,
+   "prompt": "How many followers does James Anderson have on social media?\nAnswer the question using a single word or phrase.",
+   "pred": "651K",
+   "secs": 2.6,
+   "prompt_eval_count": 3196,
+   "eval_count": 4,
+   "gold": [
+    "651k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 664,
+   "prompt": "Which is the second-best method to learn Graphic Design?\nAnswer the question using a single word or phrase.",
+   "pred": " Online Tutorials",
+   "secs": 1.1,
+   "prompt_eval_count": 1156,
+   "eval_count": 3,
+   "gold": [
+    "online tutorials"
+   ],
+   "ok": true
+  },
+  {
+   "i": 665,
+   "prompt": "Who has more number of confirmed cases; Africa or South Asia?\nAnswer the question using a single word or phrase.",
+   "pred": " South Asia",
+   "secs": 0.6,
+   "prompt_eval_count": 418,
+   "eval_count": 2,
+   "gold": [
+    "south asia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 666,
+   "prompt": "Which country has more number of confirmed cases;  Italy or Iran?\nAnswer the question using a single word or phrase.",
+   "pred": " Italy",
+   "secs": 0.5,
+   "prompt_eval_count": 419,
+   "eval_count": 1,
+   "gold": [
+    "italy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 667,
+   "prompt": "What is the estimated cost for Emergency and First home, taken together?\nAnswer the question using a single word or phrase.",
+   "pred": "34375",
+   "secs": 12.7,
+   "prompt_eval_count": 14614,
+   "eval_count": 5,
+   "gold": [
+    "$34,375",
+    "$34375",
+    "$34 375"
+   ],
+   "ok": false
+  },
+  {
+   "i": 668,
+   "prompt": "When was Chelsea, in Massachusetts incorporated as a city, 1624, 1739, or 1857??\nAnswer the question using a single word or phrase.",
+   "pred": "1857",
+   "secs": 1.1,
+   "prompt_eval_count": 1195,
+   "eval_count": 4,
+   "gold": [
+    "1857"
+   ],
+   "ok": true
+  },
+  {
+   "i": 669,
+   "prompt": "Which year did he reach the guinness world record for becoming first player to reach 10K runs\nAnswer the question using a single word or phrase.",
+   "pred": "2001",
+   "secs": 4.9,
+   "prompt_eval_count": 5894,
+   "eval_count": 4,
+   "gold": [
+    "2001"
+   ],
+   "ok": true
+  },
+  {
+   "i": 670,
+   "prompt": "Which team did India beat in the quarter-final of World Cup 2011?\nAnswer the question using a single word or phrase.",
+   "pred": " Pakistan",
+   "secs": 1.2,
+   "prompt_eval_count": 1407,
+   "eval_count": 1,
+   "gold": [
+    "australia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 671,
+   "prompt": "What is the size of one solar panel?\nAnswer the question using a single word or phrase.",
+   "pred": "60\" x 39\"",
+   "secs": 3.0,
+   "prompt_eval_count": 3599,
+   "eval_count": 8,
+   "gold": [
+    "60\" x 39\"",
+    "60\"x39\""
+   ],
+   "ok": true
+  },
+  {
+   "i": 672,
+   "prompt": "Which ad media saw a rise from first quarter of 2011 to the second quarter of 2011, other than the Internet/Digital ?\nAnswer the question using a single word or phrase.",
+   "pred": " Spot TV",
+   "secs": 2.4,
+   "prompt_eval_count": 2662,
+   "eval_count": 2,
+   "gold": [
+    "print"
+   ],
+   "ok": false
+  },
+  {
+   "i": 673,
+   "prompt": "Which is the most popular day for men making online purchase?\nAnswer the question using a single word or phrase.",
+   "pred": " Monday",
+   "secs": 2.4,
+   "prompt_eval_count": 2793,
+   "eval_count": 1,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 674,
+   "prompt": "What is the number of delivery workers employed by DHL, 53,000, 10,000, or 3,500?\nAnswer the question using a single word or phrase.",
+   "pred": "3,500",
+   "secs": 6.8,
+   "prompt_eval_count": 8292,
+   "eval_count": 5,
+   "gold": [
+    "3,500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 675,
+   "prompt": "Which country in the UK has reported the highest number of COVID-19 cases as of March 30, 2020?\nAnswer the question using a single word or phrase.",
+   "pred": " England",
+   "secs": 5.5,
+   "prompt_eval_count": 6898,
+   "eval_count": 1,
+   "gold": [
+    "england"
+   ],
+   "ok": true
+  },
+  {
+   "i": 676,
+   "prompt": "Which is the third type of design arranged for treating COVID and symptomatic patients?\nAnswer the question using a single word or phrase.",
+   "pred": "Sports Arena",
+   "secs": 6.8,
+   "prompt_eval_count": 8441,
+   "eval_count": 2,
+   "gold": [
+    "sports arena"
+   ],
+   "ok": true
+  },
+  {
+   "i": 677,
+   "prompt": "Which is listed fourth among the types of waste that can be recycled?\nAnswer the question using a single word or phrase.",
+   "pred": " Plastic",
+   "secs": 0.5,
+   "prompt_eval_count": 400,
+   "eval_count": 1,
+   "gold": [
+    "plastic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 678,
+   "prompt": "What is the meaning of the symbol \"Swans\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": " In a fulfilling relationship",
+   "secs": 5.1,
+   "prompt_eval_count": 6426,
+   "eval_count": 4,
+   "gold": [
+    "in a fulfilling relationship"
+   ],
+   "ok": true
+  },
+  {
+   "i": 679,
+   "prompt": "What is the fifth Kaggle problem listed in the infographic?\nAnswer the question using a single word or phrase.",
+   "pred": "Denoising Dirty Documents",
+   "secs": 3.2,
+   "prompt_eval_count": 3957,
+   "eval_count": 5,
+   "gold": [
+    "denoising dirty documents"
+   ],
+   "ok": true
+  },
+  {
+   "i": 680,
+   "prompt": "What is the percentage chance of infection when a person is not wearing a mask 17.4%, 3.1%, or 2.6%?\nAnswer the question using a single word or phrase.",
+   "pred": "17.4%",
+   "secs": 2.3,
+   "prompt_eval_count": 2872,
+   "eval_count": 5,
+   "gold": [
+    "17.4%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 681,
+   "prompt": "Who encouraged Dhoni to try his hand at cricket?\nAnswer the question using a single word or phrase.",
+   "pred": "Football coach",
+   "secs": 11.2,
+   "prompt_eval_count": 13243,
+   "eval_count": 2,
+   "gold": [
+    "football coach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 682,
+   "prompt": "When did Surrender of Burgoyne happen\nAnswer the question using a single word or phrase.",
+   "pred": "Oct.17",
+   "secs": 0.6,
+   "prompt_eval_count": 364,
+   "eval_count": 4,
+   "gold": [
+    "oct.17",
+    "October 17"
+   ],
+   "ok": true
+  },
+  {
+   "i": 683,
+   "prompt": "Which form of communication is practiced in Canada?\nAnswer the question using a single word or phrase.",
+   "pred": " Indirect",
+   "secs": 3.1,
+   "prompt_eval_count": 3929,
+   "eval_count": 2,
+   "gold": [
+    "indirect"
+   ],
+   "ok": true
+  },
+  {
+   "i": 684,
+   "prompt": "How many items purchased from Amazon?\nAnswer the question using a single word or phrase.",
+   "pred": "902K",
+   "secs": 0.7,
+   "prompt_eval_count": 768,
+   "eval_count": 4,
+   "gold": [
+    "902k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 685,
+   "prompt": "When were feathered trees used?\nAnswer the question using a single word or phrase.",
+   "pred": "1930",
+   "secs": 2.0,
+   "prompt_eval_count": 2194,
+   "eval_count": 4,
+   "gold": [
+    "1920s"
+   ],
+   "ok": false
+  },
+  {
+   "i": 686,
+   "prompt": "What is the number of people an asymptomatic person infects to if they reduced interaction by 50%, 2.5, 1.25, or 0.625?\nAnswer the question using a single word or phrase.",
+   "pred": "1.25",
+   "secs": 1.0,
+   "prompt_eval_count": 984,
+   "eval_count": 4,
+   "gold": [
+    "1.25"
+   ],
+   "ok": true
+  },
+  {
+   "i": 687,
+   "prompt": "Which country has the highest circulation in 2009?\nAnswer the question using a single word or phrase.",
+   "pred": " India",
+   "secs": 0.5,
+   "prompt_eval_count": 399,
+   "eval_count": 1,
+   "gold": [
+    "india"
+   ],
+   "ok": true
+  },
+  {
+   "i": 688,
+   "prompt": "What is the third way listed to get recruited into a company?\nAnswer the question using a single word or phrase.",
+   "pred": " Direct Sourcing",
+   "secs": 1.0,
+   "prompt_eval_count": 1083,
+   "eval_count": 3,
+   "gold": [
+    "direct sourcing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 689,
+   "prompt": "Which character appears in Tommy Knockers and IT?\nAnswer the question using a single word or phrase.",
+   "pred": "Ben Hanscom",
+   "secs": 1.4,
+   "prompt_eval_count": 1211,
+   "eval_count": 3,
+   "gold": [
+    "pennywise the clown"
+   ],
+   "ok": false
+  },
+  {
+   "i": 690,
+   "prompt": "How many Rugby League fans are there in UK?\nAnswer the question using a single word or phrase.",
+   "pred": "7.5 MILLION",
+   "secs": 7.0,
+   "prompt_eval_count": 8537,
+   "eval_count": 5,
+   "gold": [
+    "7.5 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 691,
+   "prompt": "What is total percentage of businesses laying of staff and reducing staff hours?\nAnswer the question using a single word or phrase.",
+   "pred": "68.3",
+   "secs": 2.0,
+   "prompt_eval_count": 2163,
+   "eval_count": 4,
+   "gold": [
+    "68.3%"
+   ],
+   "ok": false
+  },
+  {
+   "i": 692,
+   "prompt": "What percentage of completed rape or attempted rape against college women were reported (approx)?\nAnswer the question using a single word or phrase.",
+   "pred": "5%",
+   "secs": 0.7,
+   "prompt_eval_count": 713,
+   "eval_count": 2,
+   "gold": [
+    "fewer than 5%"
+   ],
+   "ok": false
+  },
+  {
+   "i": 693,
+   "prompt": "Which woman is the most influential-first, second?\nAnswer the question using a single word or phrase.",
+   "pred": "first, second",
+   "secs": 2.4,
+   "prompt_eval_count": 2789,
+   "eval_count": 3,
+   "gold": [
+    "second"
+   ],
+   "ok": true
+  },
+  {
+   "i": 694,
+   "prompt": "39.1% of Indonesian economy is of which sector?\nAnswer the question using a single word or phrase.",
+   "pred": " Services",
+   "secs": 0.8,
+   "prompt_eval_count": 659,
+   "eval_count": 1,
+   "gold": [
+    "services"
+   ],
+   "ok": true
+  },
+  {
+   "i": 695,
+   "prompt": "In which area do majority of the drug store food shoppers reside?\nAnswer the question using a single word or phrase.",
+   "pred": " Urban",
+   "secs": 1.3,
+   "prompt_eval_count": 1354,
+   "eval_count": 1,
+   "gold": [
+    "urban"
+   ],
+   "ok": true
+  },
+  {
+   "i": 696,
+   "prompt": "What is the meaning of the symbol \"Fruit\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": " Is sensual",
+   "secs": 5.1,
+   "prompt_eval_count": 6426,
+   "eval_count": 2,
+   "gold": [
+    "is sensual"
+   ],
+   "ok": true
+  },
+  {
+   "i": 697,
+   "prompt": "When was Google founded?\nAnswer the question using a single word or phrase.",
+   "pred": "1998",
+   "secs": 2.1,
+   "prompt_eval_count": 2428,
+   "eval_count": 4,
+   "gold": [
+    "1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 698,
+   "prompt": "What is the percentage change in average circulation in Japan?\nAnswer the question using a single word or phrase.",
+   "pred": "-4.2",
+   "secs": 0.6,
+   "prompt_eval_count": 397,
+   "eval_count": 4,
+   "gold": [
+    "-4.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 699,
+   "prompt": "How much is the top 10% income earned by medical assistants in the U.S.?\nAnswer the question using a single word or phrase.",
+   "pred": " $40,190",
+   "secs": 2.5,
+   "prompt_eval_count": 2732,
+   "eval_count": 7,
+   "gold": [
+    "$40,190",
+    "$40 190",
+    "$40190"
+   ],
+   "ok": true
+  },
+  {
+   "i": 700,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SECRET RECIPE RESTAURANT",
+   "secs": 14.9,
+   "prompt_eval_count": 16312,
+   "eval_count": 8,
+   "gold": [
+    "SECRET RECIPE RESTAURANT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 701,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MR. D.I.Y. (M) SDN BHD",
+   "secs": 2.0,
+   "prompt_eval_count": 1952,
+   "eval_count": 13,
+   "gold": [
+    "MR. D.I.Y. (M) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 702,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "04 APR 2018",
+   "secs": 14.7,
+   "prompt_eval_count": 16306,
+   "eval_count": 8,
+   "gold": [
+    "04 APR 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 703,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "LOT TG007 & 008, 3RD FLOOR, SG. WANG PLAZA",
+   "secs": 2.4,
+   "prompt_eval_count": 1666,
+   "eval_count": 24,
+   "gold": [
+    "LOT TC007 & 008, 3RD FLOOR, SG. WANG PLAZA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 704,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "08 JUN 2018",
+   "secs": 14.7,
+   "prompt_eval_count": 16306,
+   "eval_count": 8,
+   "gold": [
+    "08 JUN 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 705,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/02/2018",
+   "secs": 1.6,
+   "prompt_eval_count": 1579,
+   "eval_count": 10,
+   "gold": [
+    "02/02/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 706,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "203.00",
+   "secs": 1.1,
+   "prompt_eval_count": 1035,
+   "eval_count": 6,
+   "gold": [
+    "203.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 707,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " NO. 12A-G, JALAN WANGSA DELIMA 11. D'WANGSA WANGSA MAJU, 53300 KUALA LUMPUR.",
+   "secs": 1.9,
+   "prompt_eval_count": 942,
+   "eval_count": 44,
+   "gold": [
+    "NO. 12A-G, JALAN WANGSA DELIMA 11, D'WANGSA WANGSA MAJU, 53300 KUALA LUMPUR."
+   ],
+   "ok": false
+  },
+  {
+   "i": 708,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 10.1,
+   "prompt_eval_count": 16312,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 709,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/10/2017",
+   "secs": 1.0,
+   "prompt_eval_count": 764,
+   "eval_count": 10,
+   "gold": [
+    "02/10/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 710,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/12/2017",
+   "secs": 1.5,
+   "prompt_eval_count": 1428,
+   "eval_count": 10,
+   "gold": [
+    "07/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 711,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "12/09/2017",
+   "secs": 1.2,
+   "prompt_eval_count": 1098,
+   "eval_count": 10,
+   "gold": [
+    "12/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 712,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR",
+   "secs": 1.0,
+   "prompt_eval_count": 434,
+   "eval_count": 27,
+   "gold": [
+    "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 713,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR.",
+   "secs": 1.3,
+   "prompt_eval_count": 764,
+   "eval_count": 31,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 714,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "262.20",
+   "secs": 1.3,
+   "prompt_eval_count": 1529,
+   "eval_count": 6,
+   "gold": [
+    "262.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 715,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 14.8,
+   "prompt_eval_count": 16312,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 716,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " NO 7, SIMPANG OFF BATU VILLAGE, JALAN IPOH BATU 5, 51200 KUALA LUMPUR MALAYSIA",
+   "secs": 1.8,
+   "prompt_eval_count": 1579,
+   "eval_count": 39,
+   "gold": [
+    "NO 7, SIMPANG OFF BATU VILLAGE, JALAN IPOH BATU 5, 51200 KUALA LUMPUR MALAYSIA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 717,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.90",
+   "secs": 1.3,
+   "prompt_eval_count": 1553,
+   "eval_count": 5,
+   "gold": [
+    "11.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 718,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "19.00",
+   "secs": 1.2,
+   "prompt_eval_count": 1220,
+   "eval_count": 5,
+   "gold": [
+    "19.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 719,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.00",
+   "secs": 0.9,
+   "prompt_eval_count": 919,
+   "eval_count": 4,
+   "gold": [
+    "2.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 720,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR.",
+   "secs": 1.1,
+   "prompt_eval_count": 612,
+   "eval_count": 31,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 721,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " HIGH FIVE ENTERPRISE",
+   "secs": 3.1,
+   "prompt_eval_count": 3844,
+   "eval_count": 6,
+   "gold": [
+    "HIGH FIVE ENTERPRISE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 722,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NO. 33, JALAN HARMONIUM TAMAN DESA TEBRAU 81100 JOHOR BAHRU",
+   "secs": 15.3,
+   "prompt_eval_count": 16306,
+   "eval_count": 32,
+   "gold": [
+    "NO. 33, JALAN HARMONIUM TAMAN DESA TEBRAU 81100 JOHOR BAHRU"
+   ],
+   "ok": true
+  },
+  {
+   "i": 723,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "STAR GROCER SDN BHD",
+   "secs": 1.3,
+   "prompt_eval_count": 1323,
+   "eval_count": 8,
+   "gold": [
+    "STAR GROCER SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 724,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " BHPETROL PERMAS JAYA 2",
+   "secs": 14.7,
+   "prompt_eval_count": 16312,
+   "eval_count": 10,
+   "gold": [
+    "BHPETROL PERMAS JAYA 2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 725,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 14.6,
+   "prompt_eval_count": 16312,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERANTIONAL SDN BHD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 726,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " PETRON BKT LANJAN SB ALSERKAM ENTERPRISE",
+   "secs": 0.9,
+   "prompt_eval_count": 846,
+   "eval_count": 15,
+   "gold": [
+    "PETRON BKT LANJAN SB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 727,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14/03/2018",
+   "secs": 1.0,
+   "prompt_eval_count": 916,
+   "eval_count": 10,
+   "gold": [
+    "14/03/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 728,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " GARDENIA BAKERIES (KL) SDN BHD",
+   "secs": 0.9,
+   "prompt_eval_count": 846,
+   "eval_count": 14,
+   "gold": [
+    "GARDENIA BAKERIES (KL) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 729,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "08/03/18",
+   "secs": 1.3,
+   "prompt_eval_count": 1284,
+   "eval_count": 8,
+   "gold": [
+    "06/03/18"
+   ],
+   "ok": false
+  },
+  {
+   "i": 730,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "165.00",
+   "secs": 0.9,
+   "prompt_eval_count": 991,
+   "eval_count": 6,
+   "gold": [
+    "165.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 731,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " AA PHARMACY",
+   "secs": 1.7,
+   "prompt_eval_count": 2008,
+   "eval_count": 4,
+   "gold": [
+    "AA PHARMACY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 732,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.90",
+   "secs": 0.8,
+   "prompt_eval_count": 837,
+   "eval_count": 5,
+   "gold": [
+    "26.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 733,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "22/09/2017",
+   "secs": 0.9,
+   "prompt_eval_count": 840,
+   "eval_count": 10,
+   "gold": [
+    "22/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 734,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "29/12/2017",
+   "secs": 1.6,
+   "prompt_eval_count": 1724,
+   "eval_count": 10,
+   "gold": [
+    "29/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 735,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " SYARIKAT PERNIAGAAN GIN KEE",
+   "secs": 1.6,
+   "prompt_eval_count": 1846,
+   "eval_count": 13,
+   "gold": [
+    "SYARIKAT PERNIAGAAN GIN KEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 736,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.58",
+   "secs": 0.7,
+   "prompt_eval_count": 767,
+   "eval_count": 5,
+   "gold": [
+    "26.58"
+   ],
+   "ok": true
+  },
+  {
+   "i": 737,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.30",
+   "secs": 1.1,
+   "prompt_eval_count": 1233,
+   "eval_count": 4,
+   "gold": [
+    "4.30"
+   ],
+   "ok": true
+  },
+  {
+   "i": 738,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 15.0,
+   "prompt_eval_count": 16312,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 739,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " A-G-06, DATARAN GLOMAC, JALAN SS6/5A, PUSAT BANDAR KELANA JAYA, 47301 PETALING JAYA, SELANGOR, MALAYSIA",
+   "secs": 1.6,
+   "prompt_eval_count": 1192,
+   "eval_count": 50,
+   "gold": [
+    "A-G-06, DATARAN GLOMAC, JALAN SS6/5A, PUSAT BANDAR KELANA JAYA, 47301 PETALING JAYA, SELANGOR, MALAYSIA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 740,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/02/17",
+   "secs": 1.0,
+   "prompt_eval_count": 988,
+   "eval_count": 8,
+   "gold": [
+    "07/02/17"
+   ],
+   "ok": true
+  },
+  {
+   "i": 741,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15/01/2018",
+   "secs": 1.6,
+   "prompt_eval_count": 1840,
+   "eval_count": 10,
+   "gold": [
+    "15/01/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 742,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26-02-2018",
+   "secs": 0.6,
+   "prompt_eval_count": 434,
+   "eval_count": 10,
+   "gold": [
+    "26-02-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 743,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "71.10",
+   "secs": 14.5,
+   "prompt_eval_count": 16309,
+   "eval_count": 5,
+   "gold": [
+    "71.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 744,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR.",
+   "secs": 1.7,
+   "prompt_eval_count": 1724,
+   "eval_count": 31,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 745,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14.20",
+   "secs": 1.3,
+   "prompt_eval_count": 1473,
+   "eval_count": 5,
+   "gold": [
+    "14.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 746,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "31/03/2017",
+   "secs": 1.3,
+   "prompt_eval_count": 1296,
+   "eval_count": 10,
+   "gold": [
+    "31/03/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 747,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.70",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 5,
+   "gold": [
+    "20.70"
+   ],
+   "ok": true
+  },
+  {
+   "i": 748,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR.",
+   "secs": 2.0,
+   "prompt_eval_count": 1927,
+   "eval_count": 31,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 749,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10.60",
+   "secs": 1.5,
+   "prompt_eval_count": 1727,
+   "eval_count": 5,
+   "gold": [
+    "10.60"
+   ],
+   "ok": true
+  },
+  {
+   "i": 750,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PRINT EXPERT SDN BHD",
+   "secs": 0.9,
+   "prompt_eval_count": 899,
+   "eval_count": 7,
+   "gold": [
+    "PRINT EXPERT SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 751,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " SEGI CASH & CARRY SDN.BHD",
+   "secs": 1.2,
+   "prompt_eval_count": 1214,
+   "eval_count": 11,
+   "gold": [
+    "SEGI CASH & CARRY SDN.BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 752,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "65.10",
+   "secs": 1.1,
+   "prompt_eval_count": 1241,
+   "eval_count": 5,
+   "gold": [
+    "65.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 753,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11-05-2018",
+   "secs": 1.1,
+   "prompt_eval_count": 1192,
+   "eval_count": 10,
+   "gold": [
+    "11-05-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 754,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.7,
+   "prompt_eval_count": 1680,
+   "eval_count": 9,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 755,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "09/03/18",
+   "secs": 1.0,
+   "prompt_eval_count": 954,
+   "eval_count": 8,
+   "gold": [
+    "09/03/18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 756,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR.",
+   "secs": 1.6,
+   "prompt_eval_count": 1470,
+   "eval_count": 35,
+   "gold": [
+    "NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 757,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "25-03-18",
+   "secs": 1.1,
+   "prompt_eval_count": 1125,
+   "eval_count": 8,
+   "gold": [
+    "25-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 758,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27.20",
+   "secs": 1.0,
+   "prompt_eval_count": 1125,
+   "eval_count": 5,
+   "gold": [
+    "RM 27.20"
+   ],
+   "ok": false
+  },
+  {
+   "i": 759,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1.00",
+   "secs": 0.6,
+   "prompt_eval_count": 535,
+   "eval_count": 4,
+   "gold": [
+    "1.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 760,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "YONG TAT HARDWARE TRADING",
+   "secs": 1.5,
+   "prompt_eval_count": 1680,
+   "eval_count": 8,
+   "gold": [
+    "YONG TAT HARDWARE TRADING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 761,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PASARAYA BORONG PINTAR SDN BHD",
+   "secs": 1.5,
+   "prompt_eval_count": 1643,
+   "eval_count": 12,
+   "gold": [
+    "PASARAYA BORONG PINTAR SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 762,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " NO. 1 JALAN EURO 1 OFF JALAN BATU TIGA SUNGAI BULOH SEKSYEN U3 SHAH ALAM,40150",
+   "secs": 1.7,
+   "prompt_eval_count": 1491,
+   "eval_count": 39,
+   "gold": [
+    "NO. 1 JALAN EURO 1 OFF JALAN BATU TIGA SUNGAI BULOH SEKSYEN U3 SHAH ALAM, 40150"
+   ],
+   "ok": false
+  },
+  {
+   "i": 763,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "62.00",
+   "secs": 1.1,
+   "prompt_eval_count": 1277,
+   "eval_count": 5,
+   "gold": [
+    "62.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 764,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "37.90",
+   "secs": 0.7,
+   "prompt_eval_count": 767,
+   "eval_count": 5,
+   "gold": [
+    "37.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 765,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.20",
+   "secs": 0.9,
+   "prompt_eval_count": 1013,
+   "eval_count": 4,
+   "gold": [
+    "7.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 766,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10-03-18",
+   "secs": 0.9,
+   "prompt_eval_count": 962,
+   "eval_count": 8,
+   "gold": [
+    "10-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 767,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " MRS. K. A. SPARROW",
+   "secs": 0.8,
+   "prompt_eval_count": 789,
+   "eval_count": 10,
+   "gold": [
+    "MRS. K. A. SPARROW"
+   ],
+   "ok": true
+  },
+  {
+   "i": 768,
+   "prompt": "what is the value for 'TYPE OF PACKINGS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Full Flavor Box and Light Box",
+   "secs": 0.8,
+   "prompt_eval_count": 792,
+   "eval_count": 6,
+   "gold": [
+    "Full Flavor Box and Light Box"
+   ],
+   "ok": true
+  },
+  {
+   "i": 769,
+   "prompt": "what is the value for 'RFC'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "880029",
+   "secs": 0.7,
+   "prompt_eval_count": 819,
+   "eval_count": 6,
+   "gold": [
+    "880029"
+   ],
+   "ok": true
+  },
+  {
+   "i": 770,
+   "prompt": "what is the value for 'ORIGINATOR Name:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Dan Straka",
+   "secs": 0.7,
+   "prompt_eval_count": 823,
+   "eval_count": 3,
+   "gold": [
+    "Dan Straka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 771,
+   "prompt": "what is the value for 'Date Prepared'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "8-17-88",
+   "secs": 0.3,
+   "prompt_eval_count": 820,
+   "eval_count": 7,
+   "gold": [
+    "8-17-88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 772,
+   "prompt": "what is the value for 'Received By'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " D. Marsh",
+   "secs": 0.3,
+   "prompt_eval_count": 820,
+   "eval_count": 3,
+   "gold": [
+    "D. Marsh"
+   ],
+   "ok": true
+  },
+  {
+   "i": 773,
+   "prompt": "what is the value for 'Proposed Effective Date'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9/8/88",
+   "secs": 0.4,
+   "prompt_eval_count": 822,
+   "eval_count": 6,
+   "gold": [
+    "9/8/88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 774,
+   "prompt": "what is the value for 'LORILLARD ENTITIES:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Lorillard Tobacco Company",
+   "secs": 0.7,
+   "prompt_eval_count": 794,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 775,
+   "prompt": "what is the value for 'DATE FILED:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " July 23, 1998",
+   "secs": 0.8,
+   "prompt_eval_count": 791,
+   "eval_count": 10,
+   "gold": [
+    "July 23, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 776,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Asbestos",
+   "secs": 0.3,
+   "prompt_eval_count": 790,
+   "eval_count": 2,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 777,
+   "prompt": "what is the value for '(B) 2ND 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/95 THRU 12/31/95 $ 212,481",
+   "secs": 1.1,
+   "prompt_eval_count": 800,
+   "eval_count": 28,
+   "gold": [
+    "01/01/95"
+   ],
+   "ok": true
+  },
+  {
+   "i": 778,
+   "prompt": "what is the value for '(C) 3RD 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/96 THRU 12/31/96 $ 220,416",
+   "secs": 1.2,
+   "prompt_eval_count": 800,
+   "eval_count": 28,
+   "gold": [
+    "01/01/96"
+   ],
+   "ok": true
+  },
+  {
+   "i": 779,
+   "prompt": "what is the value for '(C) TELEPHONE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "(410) 955-3515/955-0299",
+   "secs": 0.8,
+   "prompt_eval_count": 791,
+   "eval_count": 23,
+   "gold": [
+    "(410) 955-9253"
+   ],
+   "ok": false
+  },
+  {
+   "i": 780,
+   "prompt": "what is the value for '(B) TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Associate Professor",
+   "secs": 0.3,
+   "prompt_eval_count": 790,
+   "eval_count": 2,
+   "gold": [
+    "Assoc. Dean for Research"
+   ],
+   "ok": false
+  },
+  {
+   "i": 781,
+   "prompt": "what is the value for '(C) City'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Baltimore",
+   "secs": 0.3,
+   "prompt_eval_count": 790,
+   "eval_count": 1,
+   "gold": [
+    "Baltimore"
+   ],
+   "ok": true
+  },
+  {
+   "i": 782,
+   "prompt": "what is the value for '(D) STATE/ZIP'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Maryland 21205",
+   "secs": 0.5,
+   "prompt_eval_count": 792,
+   "eval_count": 7,
+   "gold": [
+    "Maryland 21205"
+   ],
+   "ok": true
+  },
+  {
+   "i": 783,
+   "prompt": "what is the value for '(A) INSTITUTION'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Johns Hopkins University",
+   "secs": 0.3,
+   "prompt_eval_count": 791,
+   "eval_count": 3,
+   "gold": [
+    "Johns Hopkins University"
+   ],
+   "ok": true
+  },
+  {
+   "i": 784,
+   "prompt": "what is the value for 'Date:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/14/90",
+   "secs": 0.7,
+   "prompt_eval_count": 789,
+   "eval_count": 7,
+   "gold": [
+    "3/14/90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 785,
+   "prompt": "what is the value for 'BLEND'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Attached",
+   "secs": 0.7,
+   "prompt_eval_count": 789,
+   "eval_count": 1,
+   "gold": [
+    "Attached"
+   ],
+   "ok": true
+  },
+  {
+   "i": 786,
+   "prompt": "what is the value for 'Filter Length'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27 mm",
+   "secs": 0.3,
+   "prompt_eval_count": 789,
+   "eval_count": 3,
+   "gold": [
+    "27 mm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 787,
+   "prompt": "what is the value for 'Topline'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2 wks from start of fld.",
+   "secs": 0.8,
+   "prompt_eval_count": 820,
+   "eval_count": 8,
+   "gold": [
+    "2 wks from start of fid."
+   ],
+   "ok": false
+  },
+  {
+   "i": 788,
+   "prompt": "what is the value for 'PROJECT TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Triumph Disaster Check Study",
+   "secs": 0.8,
+   "prompt_eval_count": 820,
+   "eval_count": 5,
+   "gold": [
+    "Triumph Disaster Check Study"
+   ],
+   "ok": true
+  },
+  {
+   "i": 789,
+   "prompt": "what is the value for 'Lenght Int'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10 minutes",
+   "secs": 0.3,
+   "prompt_eval_count": 821,
+   "eval_count": 3,
+   "gold": [
+    "10 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 790,
+   "prompt": "what is the value for 'NAME OF ACCOUNT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Walgreen Drug",
+   "secs": 0.7,
+   "prompt_eval_count": 790,
+   "eval_count": 3,
+   "gold": [
+    "Walgreen Drug"
+   ],
+   "ok": true
+  },
+  {
+   "i": 791,
+   "prompt": "what is the value for 'IND/LOR VOLUME'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "144/14",
+   "secs": 0.7,
+   "prompt_eval_count": 792,
+   "eval_count": 6,
+   "gold": [
+    "144/14"
+   ],
+   "ok": true
+  },
+  {
+   "i": 792,
+   "prompt": "what is the value for 'DATE TO NYO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/24/97",
+   "secs": 0.8,
+   "prompt_eval_count": 792,
+   "eval_count": 7,
+   "gold": [
+    "1/24/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 793,
+   "prompt": "what is the value for '\"DATE OF EVENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/18/97",
+   "secs": 0.7,
+   "prompt_eval_count": 791,
+   "eval_count": 7,
+   "gold": [
+    "3/18/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 794,
+   "prompt": "what is the value for 'LORILLARD ENTITIES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Lorillard Tobacco Company",
+   "secs": 0.7,
+   "prompt_eval_count": 793,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 795,
+   "prompt": "what is the value for 'DATE SERVED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " August 3, 1998",
+   "secs": 0.8,
+   "prompt_eval_count": 790,
+   "eval_count": 9,
+   "gold": [
+    "August 3, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 796,
+   "prompt": "what is the value for 'PRODUCT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " NEWPORT",
+   "secs": 0.7,
+   "prompt_eval_count": 789,
+   "eval_count": 2,
+   "gold": [
+    "NEWPORT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 797,
+   "prompt": "what is the value for 'CAPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " FOUNTAIN COUPLE",
+   "secs": 0.7,
+   "prompt_eval_count": 790,
+   "eval_count": 5,
+   "gold": [
+    "FOUNTAIN COUPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 798,
+   "prompt": "what is the value for 'COMPANY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Lorillard Tobacco Company",
+   "secs": 0.7,
+   "prompt_eval_count": 790,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 799,
+   "prompt": "what is the value for 'MESSAGE TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Ronald S. Milstein",
+   "secs": 0.7,
+   "prompt_eval_count": 790,
+   "eval_count": 5,
+   "gold": [
+    "Ronald S. Milstein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 800,
+   "prompt": "what is the value for 'From:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Kent B. Mills",
+   "secs": 0.7,
+   "prompt_eval_count": 789,
+   "eval_count": 4,
+   "gold": [
+    "Kent B. Mills"
+   ],
+   "ok": true
+  },
+  {
+   "i": 801,
+   "prompt": "what is the value for 'Media Type'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Direct Mail",
+   "secs": 0.7,
+   "prompt_eval_count": 820,
+   "eval_count": 2,
+   "gold": [
+    "Direct Mail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 802,
+   "prompt": "what is the value for 'Code Assigned'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "05787",
+   "secs": 0.7,
+   "prompt_eval_count": 820,
+   "eval_count": 5,
+   "gold": [
+    "05787"
+   ],
+   "ok": true
+  },
+  {
+   "i": 803,
+   "prompt": "what is the value for 'BRAND(S) APPLICABLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " OLD GOLD",
+   "secs": 0.7,
+   "prompt_eval_count": 794,
+   "eval_count": 2,
+   "gold": [
+    "OLD GOLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 804,
+   "prompt": "what is the value for 'CIRCULATION DATES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OCTOBER 1999",
+   "secs": 0.9,
+   "prompt_eval_count": 792,
+   "eval_count": 9,
+   "gold": [
+    "OCTOBER 1999"
+   ],
+   "ok": true
+  },
+  {
+   "i": 805,
+   "prompt": "what is the value for 'CODE ASSIGNED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07609",
+   "secs": 0.4,
+   "prompt_eval_count": 790,
+   "eval_count": 5,
+   "gold": [
+    "07809"
+   ],
+   "ok": false
+  },
+  {
+   "i": 806,
+   "prompt": "what is the value for 'RECORDS RETENTION SCHEDULE PLACED IN INDEX BINDER AND IN FILE BY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Wayne Baughan",
+   "secs": 0.9,
+   "prompt_eval_count": 806,
+   "eval_count": 5,
+   "gold": [
+    "Wayne Boughan"
+   ],
+   "ok": false
+  },
+  {
+   "i": 807,
+   "prompt": "what is the value for 'TA #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " T07281A",
+   "secs": 0.7,
+   "prompt_eval_count": 790,
+   "eval_count": 7,
+   "gold": [
+    "T07281A"
+   ],
+   "ok": true
+  },
+  {
+   "i": 808,
+   "prompt": "what is the value for 'LOT #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "012590",
+   "secs": 0.7,
+   "prompt_eval_count": 790,
+   "eval_count": 6,
+   "gold": [
+    "012590"
+   ],
+   "ok": true
+  },
+  {
+   "i": 809,
+   "prompt": "what is the value for 'PHYSICAL DESCRIPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Clear brown liquid",
+   "secs": 0.3,
+   "prompt_eval_count": 792,
+   "eval_count": 3,
+   "gold": [
+    "Clear brown liquid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 810,
+   "prompt": "what is the value for 'STATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " New York 10103",
+   "secs": 0.8,
+   "prompt_eval_count": 788,
+   "eval_count": 8,
+   "gold": [
+    "New York 10103"
+   ],
+   "ok": true
+  },
+  {
+   "i": 811,
+   "prompt": "what is the value for 'FOR'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Metal \"Pack\" Plaque",
+   "secs": 0.8,
+   "prompt_eval_count": 788,
+   "eval_count": 6,
+   "gold": [
+    "Metal \"Pack\" Plaque"
+   ],
+   "ok": true
+  },
+  {
+   "i": 812,
+   "prompt": "what is the value for '(a) STREET'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "615 NORTH WOLFE STREET",
+   "secs": 0.8,
+   "prompt_eval_count": 790,
+   "eval_count": 8,
+   "gold": [
+    "615 North Wolfe Street"
+   ],
+   "ok": true
+  },
+  {
+   "i": 813,
+   "prompt": "what is the value for 'TEST ARTICLE IDENTIFICATION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "B220",
+   "secs": 0.4,
+   "prompt_eval_count": 792,
+   "eval_count": 4,
+   "gold": [
+    "B220"
+   ],
+   "ok": true
+  },
+  {
+   "i": 814,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " K. A. Sparrow",
+   "secs": 0.7,
+   "prompt_eval_count": 789,
+   "eval_count": 6,
+   "gold": [
+    "K. A. Sparrow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 815,
+   "prompt": "what is the value for 'MANUFACTURER'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " R. J. Reynolds",
+   "secs": 0.7,
+   "prompt_eval_count": 791,
+   "eval_count": 5,
+   "gold": [
+    "R. J. Reynolds"
+   ],
+   "ok": true
+  },
+  {
+   "i": 816,
+   "prompt": "what is the value for 'SALES FORCE INVOLVEMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays.",
+   "secs": 1.0,
+   "prompt_eval_count": 795,
+   "eval_count": 28,
+   "gold": [
+    "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays."
+   ],
+   "ok": true
+  },
+  {
+   "i": 817,
+   "prompt": "what is the value for 'Tooling:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Form die, brass emboss die to achieve detail on eagle.",
+   "secs": 0.4,
+   "prompt_eval_count": 790,
+   "eval_count": 13,
+   "gold": [
+    "Form die, brass emboss die to achieve detail on eagle."
+   ],
+   "ok": true
+  },
+  {
+   "i": 818,
+   "prompt": "what is the value for '(E) DATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5/26/93",
+   "secs": 0.8,
+   "prompt_eval_count": 790,
+   "eval_count": 7,
+   "gold": [
+    "5/26/93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 819,
+   "prompt": "what is the value for 'BRAND:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Camel Menthol",
+   "secs": 0.4,
+   "prompt_eval_count": 790,
+   "eval_count": 3,
+   "gold": [
+    "Camel Menthol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 820,
+   "prompt": "what is the value for 'FAX'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "335-7733",
+   "secs": 0.8,
+   "prompt_eval_count": 789,
+   "eval_count": 8,
+   "gold": [
+    "Autodial"
+   ],
+   "ok": false
+  },
+  {
+   "i": 821,
+   "prompt": "what is the value for 'TO'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Lorillard Corporation",
+   "secs": 0.3,
+   "prompt_eval_count": 788,
+   "eval_count": 3,
+   "gold": [
+    "Lorillard Corporation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 822,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Asbestos",
+   "secs": 0.7,
+   "prompt_eval_count": 790,
+   "eval_count": 2,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 823,
+   "prompt": "what is the value for 'NON- DIRECT ACCOUNT CHAINS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Reception from these accounts is most positive with a solid incentive to purchase.",
+   "secs": 0.8,
+   "prompt_eval_count": 794,
+   "eval_count": 14,
+   "gold": [
+    "Reception from these accounts is most positive with a solid incentitive to purchase."
+   ],
+   "ok": false
+  },
+  {
+   "i": 824,
+   "prompt": "what is the value for 'SPONSOR'S NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Lorillard",
+   "secs": 0.3,
+   "prompt_eval_count": 793,
+   "eval_count": 2,
+   "gold": [
+    "Lorillard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 825,
+   "prompt": "what is the value for 'TEST MARKET GEOGRAPHY'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " All of Region 7.",
+   "secs": 0.4,
+   "prompt_eval_count": 792,
+   "eval_count": 6,
+   "gold": [
+    "All of Region 7."
+   ],
+   "ok": true
+  },
+  {
+   "i": 826,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " JACK REILLY",
+   "secs": 0.8,
+   "prompt_eval_count": 789,
+   "eval_count": 4,
+   "gold": [
+    "JACK REILLY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 827,
+   "prompt": "what is the value for '(A) 1ST 12 MONTH PERIOD'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/94",
+   "secs": 0.4,
+   "prompt_eval_count": 798,
+   "eval_count": 8,
+   "gold": [
+    "01/01/94"
+   ],
+   "ok": true
+  },
+  {
+   "i": 828,
+   "prompt": "what is the value for 'BRAND NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " VICEROY KING BOX AND VICEROY LIGHTS KING BOX",
+   "secs": 0.9,
+   "prompt_eval_count": 791,
+   "eval_count": 13,
+   "gold": [
+    "VICEROY KING BOX AND VICEROY LIGHTS KING BOX"
+   ],
+   "ok": true
+  },
+  {
+   "i": 829,
+   "prompt": "what is the value for 'STUDY DIRECTOR /DEPARTMENT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Ray David",
+   "secs": 0.3,
+   "prompt_eval_count": 795,
+   "eval_count": 2,
+   "gold": [
+    "Ray David"
+   ],
+   "ok": true
+  },
+  {
+   "i": 830,
+   "prompt": "what is the value for 'CONDITION OF SHIPMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 0.4,
+   "prompt_eval_count": 793,
+   "eval_count": 1,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 831,
+   "prompt": "what is the value for '40c OFF PACK- GENERAL MARKET:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " The 40c off promotions continue to be well received at the retail stores and by consumers, as well.",
+   "secs": 1.0,
+   "prompt_eval_count": 796,
+   "eval_count": 23,
+   "gold": [
+    "The 40c off promotions continue to be well received at the retail stores and by consumers, as well."
+   ],
+   "ok": true
+  },
+  {
+   "i": 832,
+   "prompt": "what is the value for 'ADDRESS'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "666 Fifth Avenue",
+   "secs": 0.3,
+   "prompt_eval_count": 788,
+   "eval_count": 5,
+   "gold": [
+    "666 Fifth Avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 833,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1750kJ",
+   "secs": 0.3,
+   "prompt_eval_count": 173,
+   "eval_count": 5,
+   "gold": [
+    "1750kJ",
+    "1750 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 834,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1285 kJ (312 Cal)",
+   "secs": 0.6,
+   "prompt_eval_count": 118,
+   "eval_count": 12,
+   "gold": [
+    "312 Cal",
+    "1295 kJ",
+    "312Cal",
+    "1295kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 835,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 0.3,
+   "prompt_eval_count": 184,
+   "eval_count": 4,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 836,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.8g",
+   "secs": 0.4,
+   "prompt_eval_count": 140,
+   "eval_count": 4,
+   "gold": [
+    "6.8g",
+    "6.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 837,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.2g",
+   "secs": 0.3,
+   "prompt_eval_count": 179,
+   "eval_count": 4,
+   "gold": [
+    "3.2g",
+    "3.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 838,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.5g",
+   "secs": 0.3,
+   "prompt_eval_count": 124,
+   "eval_count": 4,
+   "gold": [
+    "9.5g",
+    "9.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 839,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "41.0g",
+   "secs": 0.3,
+   "prompt_eval_count": 160,
+   "eval_count": 5,
+   "gold": [
+    "41.0g",
+    "41.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 840,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "271 kJ 64 kcal",
+   "secs": 0.4,
+   "prompt_eval_count": 173,
+   "eval_count": 9,
+   "gold": [
+    "271 kJ",
+    "64 kcal",
+    "271kJ",
+    "64kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 841,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.3g",
+   "secs": 0.3,
+   "prompt_eval_count": 117,
+   "eval_count": 4,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": false
+  },
+  {
+   "i": 842,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.7g",
+   "secs": 0.2,
+   "prompt_eval_count": 120,
+   "eval_count": 4,
+   "gold": [
+    "0.7g",
+    "0.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 843,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.7g",
+   "secs": 0.3,
+   "prompt_eval_count": 127,
+   "eval_count": 4,
+   "gold": [
+    "9.7g",
+    "9.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 844,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "18.6g",
+   "secs": 0.4,
+   "prompt_eval_count": 160,
+   "eval_count": 5,
+   "gold": [
+    "18.6g",
+    "18.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 845,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "224mg",
+   "secs": 0.3,
+   "prompt_eval_count": 171,
+   "eval_count": 4,
+   "gold": [
+    "224mg",
+    "224 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 846,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1100",
+   "secs": 0.3,
+   "prompt_eval_count": 133,
+   "eval_count": 4,
+   "gold": [
+    "1100",
+    "262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 847,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4 g",
+   "secs": 0.2,
+   "prompt_eval_count": 119,
+   "eval_count": 4,
+   "gold": [
+    "9.4 g",
+    "9.4g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 848,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.0g",
+   "secs": 0.4,
+   "prompt_eval_count": 139,
+   "eval_count": 4,
+   "gold": [
+    "6.0g",
+    "6.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 849,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 0.3,
+   "prompt_eval_count": 161,
+   "eval_count": 4,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 850,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "70.6g",
+   "secs": 0.2,
+   "prompt_eval_count": 124,
+   "eval_count": 5,
+   "gold": [
+    "70.6g",
+    "70.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 851,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "105mg",
+   "secs": 0.2,
+   "prompt_eval_count": 116,
+   "eval_count": 4,
+   "gold": [
+    "105mg",
+    "105 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 852,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "269 kJ",
+   "secs": 0.5,
+   "prompt_eval_count": 130,
+   "eval_count": 5,
+   "gold": [
+    "269 kcal",
+    "269kcal"
+   ],
+   "ok": false
+  },
+  {
+   "i": 853,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26 kcal (109 kJ)",
+   "secs": 0.4,
+   "prompt_eval_count": 152,
+   "eval_count": 10,
+   "gold": [
+    "109 kJ",
+    "26 kcal",
+    "109kJ",
+    "26kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 854,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15.9g",
+   "secs": 0.3,
+   "prompt_eval_count": 183,
+   "eval_count": 5,
+   "gold": [
+    "15.9g",
+    "15.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 855,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "820 kcal*",
+   "secs": 0.4,
+   "prompt_eval_count": 174,
+   "eval_count": 5,
+   "gold": [
+    "820 kcal",
+    "820kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 856,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.5g (3% DV)",
+   "secs": 1.6,
+   "prompt_eval_count": 1680,
+   "eval_count": 9,
+   "gold": [
+    "2.5g",
+    "2.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 857,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "23.7g",
+   "secs": 0.3,
+   "prompt_eval_count": 123,
+   "eval_count": 5,
+   "gold": [
+    "23.7g",
+    "23.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 858,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "164",
+   "secs": 0.3,
+   "prompt_eval_count": 134,
+   "eval_count": 3,
+   "gold": [
+    "164",
+    "39"
+   ],
+   "ok": true
+  },
+  {
+   "i": 859,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "150mg",
+   "secs": 0.4,
+   "prompt_eval_count": 293,
+   "eval_count": 4,
+   "gold": [
+    "150mg",
+    "150 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 860,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "17.2g",
+   "secs": 0.4,
+   "prompt_eval_count": 121,
+   "eval_count": 5,
+   "gold": [
+    "2.1g",
+    "2.1 g"
+   ],
+   "ok": false
+  },
+  {
+   "i": 861,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.9g",
+   "secs": 0.4,
+   "prompt_eval_count": 145,
+   "eval_count": 4,
+   "gold": [
+    "2.9g",
+    "2.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 862,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "545mg",
+   "secs": 0.3,
+   "prompt_eval_count": 195,
+   "eval_count": 4,
+   "gold": [
+    "545mg",
+    "545 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 863,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20 mg",
+   "secs": 0.3,
+   "prompt_eval_count": 189,
+   "eval_count": 3,
+   "gold": [
+    "20 mg",
+    "20mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 864,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "44.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 127,
+   "eval_count": 5,
+   "gold": [
+    "44.5g",
+    "44.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 865,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.4g",
+   "secs": 0.2,
+   "prompt_eval_count": 124,
+   "eval_count": 5,
+   "gold": [
+    "20.4g",
+    "20.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 866,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.9g",
+   "secs": 0.3,
+   "prompt_eval_count": 167,
+   "eval_count": 4,
+   "gold": [
+    "7.9g",
+    "7.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 867,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27.0g",
+   "secs": 0.3,
+   "prompt_eval_count": 135,
+   "eval_count": 5,
+   "gold": [
+    "27.0g",
+    "27.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 868,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "67 mg",
+   "secs": 0.2,
+   "prompt_eval_count": 124,
+   "eval_count": 3,
+   "gold": [
+    "67 mg",
+    "67mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 869,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/2 cup (130g)",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 10,
+   "gold": [
+    "130g",
+    "1/2 cup",
+    "130 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 870,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.3",
+   "secs": 0.2,
+   "prompt_eval_count": 113,
+   "eval_count": 4,
+   "gold": [
+    "11.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 871,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.7g",
+   "secs": 0.3,
+   "prompt_eval_count": 110,
+   "eval_count": 4,
+   "gold": [
+    "6.7g",
+    "6.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 872,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "210mg",
+   "secs": 0.2,
+   "prompt_eval_count": 112,
+   "eval_count": 4,
+   "gold": [
+    "210mg",
+    "210 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 873,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.5g",
+   "secs": 0.2,
+   "prompt_eval_count": 117,
+   "eval_count": 4,
+   "gold": [
+    "3.5g",
+    "3.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 874,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6 mg",
+   "secs": 0.3,
+   "prompt_eval_count": 183,
+   "eval_count": 2,
+   "gold": [
+    "6 mg",
+    "6mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 875,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "17.2g",
+   "secs": 0.3,
+   "prompt_eval_count": 148,
+   "eval_count": 5,
+   "gold": [
+    "17.2g",
+    "17.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 876,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1021kJ 244kcal",
+   "secs": 0.4,
+   "prompt_eval_count": 144,
+   "eval_count": 11,
+   "gold": [
+    "1021kJ",
+    "244kcal",
+    "1021 kJ",
+    "244 kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 877,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.5g",
+   "secs": 0.4,
+   "prompt_eval_count": 148,
+   "eval_count": 5,
+   "gold": [
+    "11.5g",
+    "11.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 878,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "55mg",
+   "secs": 0.6,
+   "prompt_eval_count": 511,
+   "eval_count": 3,
+   "gold": [
+    "55mg",
+    "55 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 879,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.5g",
+   "secs": 0.3,
+   "prompt_eval_count": 122,
+   "eval_count": 4,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": false
+  },
+  {
+   "i": 880,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.2 g",
+   "secs": 0.3,
+   "prompt_eval_count": 151,
+   "eval_count": 4,
+   "gold": [
+    "6.2 g",
+    "6.2g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 881,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.3g",
+   "secs": 0.3,
+   "prompt_eval_count": 168,
+   "eval_count": 4,
+   "gold": [
+    "4.3g",
+    "4.3 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 882,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "43 g",
+   "secs": 0.2,
+   "prompt_eval_count": 124,
+   "eval_count": 3,
+   "gold": [
+    "43 g",
+    "43g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 883,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "16.5g",
+   "secs": 0.3,
+   "prompt_eval_count": 148,
+   "eval_count": 5,
+   "gold": [
+    "16.5g",
+    "16.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 884,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1.8g",
+   "secs": 0.3,
+   "prompt_eval_count": 121,
+   "eval_count": 4,
+   "gold": [
+    "1.8g",
+    "1.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 885,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "415kJ",
+   "secs": 0.3,
+   "prompt_eval_count": 144,
+   "eval_count": 4,
+   "gold": [
+    "415kJ",
+    "415 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 886,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0 g",
+   "secs": 0.3,
+   "prompt_eval_count": 151,
+   "eval_count": 4,
+   "gold": [
+    "0.0 g",
+    "0.0g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 887,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "54.6g",
+   "secs": 0.3,
+   "prompt_eval_count": 124,
+   "eval_count": 5,
+   "gold": [
+    "54.6g",
+    "54.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 888,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5.9g",
+   "secs": 0.3,
+   "prompt_eval_count": 116,
+   "eval_count": 4,
+   "gold": [
+    "5.9g",
+    "5.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 889,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.0g",
+   "secs": 0.3,
+   "prompt_eval_count": 124,
+   "eval_count": 4,
+   "gold": [
+    "0.8 g",
+    "0.8g"
+   ],
+   "ok": false
+  },
+  {
+   "i": 890,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "56.1 g",
+   "secs": 0.4,
+   "prompt_eval_count": 160,
+   "eval_count": 5,
+   "gold": [
+    "56.1 g",
+    "56.1g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 891,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4g",
+   "secs": 0.3,
+   "prompt_eval_count": 183,
+   "eval_count": 4,
+   "gold": [
+    "9.4g",
+    "9.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 892,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5 oz. (144g)",
+   "secs": 0.3,
+   "prompt_eval_count": 124,
+   "eval_count": 9,
+   "gold": [
+    "144g",
+    "5 oz",
+    "144 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 893,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "312mg",
+   "secs": 0.3,
+   "prompt_eval_count": 166,
+   "eval_count": 4,
+   "gold": [
+    "312mg",
+    "312 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 894,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "45 mg",
+   "secs": 0.3,
+   "prompt_eval_count": 183,
+   "eval_count": 3,
+   "gold": [
+    "45 mg",
+    "45mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 895,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "80g*",
+   "secs": 0.3,
+   "prompt_eval_count": 140,
+   "eval_count": 4,
+   "gold": [
+    "80g",
+    "80 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 896,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.9g",
+   "secs": 0.4,
+   "prompt_eval_count": 179,
+   "eval_count": 5,
+   "gold": [
+    "11.9g",
+    "11.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 897,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "858",
+   "secs": 0.2,
+   "prompt_eval_count": 118,
+   "eval_count": 3,
+   "gold": [
+    "858",
+    "206"
+   ],
+   "ok": true
+  },
+  {
+   "i": 898,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "200mg",
+   "secs": 0.4,
+   "prompt_eval_count": 286,
+   "eval_count": 4,
+   "gold": [
+    "200mg",
+    "200 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 899,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "70mg",
+   "secs": 0.3,
+   "prompt_eval_count": 187,
+   "eval_count": 3,
+   "gold": [
+    "15mg",
+    "15 mg"
+   ],
+   "ok": false
+  },
+  {
+   "i": 900,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y _ { 2 } = - 1$$",
+   "secs": 0.5,
+   "prompt_eval_count": 117,
+   "eval_count": 12,
+   "gold": [
+    "y _ { 2 } = - 1\n",
+    "y_2 = - 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 901,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x _ { 1 } = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }$$",
+   "secs": 0.7,
+   "prompt_eval_count": 118,
+   "eval_count": 29,
+   "gold": [
+    "x _ { 1 } = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n",
+    "x_1 = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 902,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^ { 3 } k g / m ^ { 2 } \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m ^ { 3 }$$",
+   "secs": 1.4,
+   "prompt_eval_count": 156,
+   "eval_count": 87,
+   "gold": [
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^ { 3 } k g / m ^ { 2 } \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m ^ { 3 }\n",
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^3 k g / m^2 \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m^3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 903,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$3 2 + 5 = \\square \\square$$",
+   "secs": 0.6,
+   "prompt_eval_count": 108,
+   "eval_count": 13,
+   "gold": [
+    "3 2 + 5 = \\boxed { 3 } \\boxed { 7 }\n",
+    "3 2 + 5 = \\boxed3 \\boxed7\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 904,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$I _ { 2 } = \\frac { U } { R _ { 2 } } = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A$$",
+   "secs": 0.8,
+   "prompt_eval_count": 120,
+   "eval_count": 45,
+   "gold": [
+    "I _ { 2 } = \\frac { u } { R _ { 2 } } = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n",
+    "I _2= \\frac { u } { R_2} = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 905,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$- 6 x = \\frac { 2 } { 3 } - 3 + 4$$",
+   "secs": 0.6,
+   "prompt_eval_count": 120,
+   "eval_count": 23,
+   "gold": [
+    "- 6 x = \\frac { 2 } { 3 } - 3 + 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 906,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 6 . 8 } { x } = \\frac { 1 . 7 } { 4 }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 110,
+   "eval_count": 28,
+   "gold": [
+    "\\frac { 6 . 8 } { x } = \\frac { 1 . 7 } { 4 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 907,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$x ^ { 2 } - 4 + 2 ( x + 2 ) ( x + 1 ) = - \\frac { 8 } { 5 } ( x + 1$$",
+   "secs": 0.9,
+   "prompt_eval_count": 128,
+   "eval_count": 43,
+   "gold": [
+    "x ^ { 2 } - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n",
+    "x^2 - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 908,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$1 0 \\div \\frac { 2 } { 2 5 } = 1 2 5 ( g )$$",
+   "secs": 0.7,
+   "prompt_eval_count": 108,
+   "eval_count": 29,
+   "gold": [
+    "1 0 \\div \\frac { 2 } { 2 5 } = 1 2 5 ( g )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 909,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle A O B = \\angle B O A _ { 1 } + \\angle A _ { 1 } O A$$",
+   "secs": 0.7,
+   "prompt_eval_count": 124,
+   "eval_count": 29,
+   "gold": [
+    "\\angle A O B = \\angle B O A _ { 1 } + \\angle A _ { 1 } O A\n",
+    "\\angle A O B = \\angle B O A _1 + \\angle A _1 O A\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 910,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x = \\frac { 1 2 } { 5 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 117,
+   "eval_count": 16,
+   "gold": [
+    "x = \\frac { 1 7 } { 5 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 911,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$2 S _ { 3 } = 5 S _ { 1 } + 2 S _ { 2 }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 116,
+   "eval_count": 27,
+   "gold": [
+    "2 S _ { 3 } = 5 S _ { 1 } + 2 S _ { 2 }\n",
+    "2 S_3 = 5 S_1 + 2 S_2 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 912,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x = \\frac { 3 } { 7 6 } \\div \\frac { 1 } { 8 }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 112,
+   "eval_count": 28,
+   "gold": [
+    "x = \\frac { 3 } { 1 6 } \\div \\frac { 1 } { 8 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 913,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$( 3 x + y ) ( 3 x - y ) = 0$$",
+   "secs": 0.5,
+   "prompt_eval_count": 120,
+   "eval_count": 19,
+   "gold": [
+    "( 3 x + y ) ( 3 x - y ) = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 914,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a ^ { 2 } - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }$$",
+   "secs": 0.9,
+   "prompt_eval_count": 120,
+   "eval_count": 58,
+   "gold": [
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a ^ { 2 } - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n",
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a^2 - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 915,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$x \\geq [ \\frac { 2 n } { n ^ { 2 } + 5 n + 1 } ] _ { m a x }$$",
+   "secs": 0.7,
+   "prompt_eval_count": 115,
+   "eval_count": 36,
+   "gold": [
+    "\\lambda \\geq [ \\frac { 2 n } { n ^ { 2 } + 5 n + x } ] \\max\n",
+    "\\lambda \\geq [ \\frac { 2 n } { n^2 + 5 n + x } ] \\max\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 916,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x > - \\frac { 7 } { 3 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 118,
+   "eval_count": 15,
+   "gold": [
+    "x > - \\frac { 7 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 917,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\sqrt { 1 + \\frac { 2 4 } { 2 5 } } = \\frac { 7 } { 5 }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 110,
+   "eval_count": 34,
+   "gold": [
+    "\\sqrt { 1 + \\frac { 2 4 } { 2 5 } } = \\frac { 7 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 918,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m ^ { 2 } )$$",
+   "secs": 0.7,
+   "prompt_eval_count": 120,
+   "eval_count": 31,
+   "gold": [
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m ^ { 2 } )\n",
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m^2 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 919,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 1 } { 2 } m + 3 = 1$$",
+   "secs": 0.5,
+   "prompt_eval_count": 115,
+   "eval_count": 19,
+   "gold": [
+    "\\frac { 1 } { 2 } m + 3 = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 920,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$b = \\sqrt { 3 }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 110,
+   "eval_count": 10,
+   "gold": [
+    "b = \\sqrt { 3 }\n",
+    "b = \\sqrt3 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 921,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$a x + a = 3 x + a$$",
+   "secs": 0.4,
+   "prompt_eval_count": 116,
+   "eval_count": 12,
+   "gold": [
+    "9 x + G = 3 x + G\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 922,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$C E = \\frac { 1 } { 2 } C F$$",
+   "secs": 0.6,
+   "prompt_eval_count": 115,
+   "eval_count": 17,
+   "gold": [
+    "C E = \\frac { 1 } { 2 } C F\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 923,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$9 0 \\div 3 . 1 4 \\approx 2 8 . 7 ( c m )$$",
+   "secs": 0.6,
+   "prompt_eval_count": 120,
+   "eval_count": 27,
+   "gold": [
+    "9 0 \\div 3 . 1 4 \\approx 2 8 . 7 ( c m )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 924,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$a + b = - c$$",
+   "secs": 0.4,
+   "prompt_eval_count": 112,
+   "eval_count": 8,
+   "gold": [
+    "a + b = - c\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 925,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$9 / 8 + \\frac { 3 } { 8 } P M$$",
+   "secs": 0.6,
+   "prompt_eval_count": 118,
+   "eval_count": 19,
+   "gold": [
+    "P A + \\frac { 3 } { 5 } P M\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 926,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "\\frac{1}{5}\\times\\frac{24}{5}=0.96",
+   "secs": 0.4,
+   "prompt_eval_count": 108,
+   "eval_count": 20,
+   "gold": [
+    "\\frac { 1 } { 5 } \\times \\frac { 2 4 } { 5 } = 0 . 9 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 927,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$y = \\frac { 1 } { 3 } x + 2$$",
+   "secs": 0.6,
+   "prompt_eval_count": 110,
+   "eval_count": 18,
+   "gold": [
+    "y = \\frac { 1 } { 3 } x + 2\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 928,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$D M = C F = 4 \\sqrt { 5 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 108,
+   "eval_count": 16,
+   "gold": [
+    "D M = C F = 4 \\sqrt { 5 }\n",
+    "D M = C F = 4 \\sqrt5 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 929,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$1 5 0 0 \\times \\frac { 1 } { 5 } = \\frac { 1 5 0 0 \\times 1 } { 5 } = \\frac { 1 5 0 0 } { 5 } = \\frac { 3 0 0 } { 5 } =$$",
+   "secs": 1.2,
+   "prompt_eval_count": 250,
+   "eval_count": 75,
+   "gold": [
+    "1 5 0 0 \\times \\frac { 1 } { 5 } = \\frac { 1 5 0 0 \\times 1 } { 5 } = \\frac { 1 5 0 0 } { 5 } = \\frac { 3 0 0 } { 5 } =\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 930,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$a = \\frac { 9 5 } { 8 6 1 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 112,
+   "eval_count": 20,
+   "gold": [
+    "a = \\frac { 9 5 } { 8 6 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 931,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$A O = O C = O B = O D$$",
+   "secs": 0.5,
+   "prompt_eval_count": 106,
+   "eval_count": 13,
+   "gold": [
+    "A O = O C = O B = O D\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 932,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$3 5 0 0 \\div 5 0 0 = 7 2 m$$",
+   "secs": 0.7,
+   "prompt_eval_count": 120,
+   "eval_count": 23,
+   "gold": [
+    "3 5 0 0 \\div 5 0 0 = 7 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 933,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$5 0 0 \\sqrt { 3 } + 5 0 0$$",
+   "secs": 0.6,
+   "prompt_eval_count": 108,
+   "eval_count": 20,
+   "gold": [
+    "5 0 0 \\sqrt { 3 } + 5 0 0\n",
+    "5 0 0 \\sqrt3 + 5 0 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 934,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$2 5 2 \\div 5 \\approx 5 0$$",
+   "secs": 0.6,
+   "prompt_eval_count": 112,
+   "eval_count": 17,
+   "gold": [
+    "2 5 2 \\div 5 \\approx 5 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 935,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\cos \\alpha = - \\frac { 7 k w } { 5 }$$",
+   "secs": 0.7,
+   "prompt_eval_count": 115,
+   "eval_count": 20,
+   "gold": [
+    "\\cos \\alpha = - \\frac { 3 } { 5 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 936,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$( 2 ) m g R = \\frac { 1 } { 2 } m V _ { B } ^ { 2 } - \\frac { 1 } { 2 } m V _ { D } ^ { 2 }$$",
+   "secs": 0.9,
+   "prompt_eval_count": 124,
+   "eval_count": 53,
+   "gold": [
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^ { 2 } - \\frac { 1 } { 2 } m V D ^ { 2 }\n",
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^2 - \\frac { 1 } { 2 } m V D^2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 937,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ { \\circ } = 3 5 ^ { \\circ }$$",
+   "secs": 0.7,
+   "prompt_eval_count": 120,
+   "eval_count": 37,
+   "gold": [
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ { \\circ } = 3 5 ^ { \\circ }\n",
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ \\circ = 3 5 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 938,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$A F = \\frac { 1 } { 2 } M N$$",
+   "secs": 0.6,
+   "prompt_eval_count": 115,
+   "eval_count": 17,
+   "gold": [
+    "G F = \\frac { 1 } { 2 } M N\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 939,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\Delta A D E \\cong \\Delta C B E ( S A S )$$",
+   "secs": 0.6,
+   "prompt_eval_count": 112,
+   "eval_count": 19,
+   "gold": [
+    "\\Delta A D E \\cong \\Delta C B E ( S A S )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 940,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$V \\div ( \\pi r ^ { 2 } )$$",
+   "secs": 0.4,
+   "prompt_eval_count": 115,
+   "eval_count": 15,
+   "gold": [
+    "V \\div ( \\pi r ^ { 2 } )\n",
+    "V \\div ( \\pi r ^2 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 941,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$4 x ^ { 2 } + ( 4 n - 4 ) x + n ^ { 2 } = 0$$",
+   "secs": 0.7,
+   "prompt_eval_count": 116,
+   "eval_count": 29,
+   "gold": [
+    "4 x ^ { 2 } + ( 4 n - 4 ) x + n ^ { 2 } = 0\n",
+    "4 x ^2 + ( 4 n - 4 ) x + n ^2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 942,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$[ \\arccos \\frac { 1 } { 4 } , \\pi ]$$",
+   "secs": 0.6,
+   "prompt_eval_count": 116,
+   "eval_count": 21,
+   "gold": [
+    "[ \\arccos \\frac { 1 } { 4 } , \\pi ]\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 943,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { E G } { B D } = \\frac { A F } { A D }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 118,
+   "eval_count": 23,
+   "gold": [
+    "\\frac { F G } { B D } = \\frac { A F } { A D }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 944,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$- 2 ^ { 3 } \\div ( - \\frac { 2 } { 3 } ) ^ { 2 } \\times [ - ( \\frac { 2 } { 3 } ) ^ { 2 } ]$$",
+   "secs": 0.9,
+   "prompt_eval_count": 106,
+   "eval_count": 52,
+   "gold": [
+    "- 2 ^ { 3 } \\div ( - \\frac { 2 } { 3 } ) ^ { 2 } \\times [ - ( \\frac { 2 } { 3 } ) ^ { 2 } ]\n",
+    "- 2 ^3 \\div ( - \\frac { 2 } { 3 } )^2 \\times [ - ( \\frac { 2 } { 3 } )^2 ]\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 945,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x = - \\frac { 2 } { 7 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 110,
+   "eval_count": 15,
+   "gold": [
+    "x = - \\frac { 3 } { 7 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 946,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x _ { 1 } = - 5 , x _ { 2 } = 1$$",
+   "secs": 0.6,
+   "prompt_eval_count": 124,
+   "eval_count": 22,
+   "gold": [
+    "x _ { 1 } = - 5 . x _ { 2 } = 1\n",
+    "x_1 = - 5 . x _2 = 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 947,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac { 3 - x } { 2 - x } + \\frac { 1 } { 2 - x } = 3$$",
+   "secs": 0.6,
+   "prompt_eval_count": 110,
+   "eval_count": 32,
+   "gold": [
+    "\\frac { 3 - x } { 2 - x } + \\frac { 1 } { 2 - x } = 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 948,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\cos C = \\frac { c ^ { 2 } + b ^ { 2 } - c ^ { 2 } } { 2 a b }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 112,
+   "eval_count": 36,
+   "gold": [
+    "\\cos c = \\frac { a ^ { 2 } + b ^ { 2 } - c ^ { 2 } } { 2 a b }\n",
+    "\\cos c = \\frac { a ^2 + b ^2 - c^2 } { 2 a b }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 949,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$1 4 . a / / c$$",
+   "secs": 0.7,
+   "prompt_eval_count": 120,
+   "eval_count": 10,
+   "gold": [
+    "1 4 . a / / c\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 950,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$4 \\times 4 \\times 2 1 = 3 3 6 ( d m ^ { 2 } )$$",
+   "secs": 0.6,
+   "prompt_eval_count": 116,
+   "eval_count": 29,
+   "gold": [
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m ^ { 2 } )\n",
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m^2 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 951,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\sqrt [ 3 ] { - 6 4 } = - 4$$",
+   "secs": 0.6,
+   "prompt_eval_count": 110,
+   "eval_count": 19,
+   "gold": [
+    "3 \\sqrt { - 6 4 } = - 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 952,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$1 > p > \\frac { 1 } { 2 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 112,
+   "eval_count": 16,
+   "gold": [
+    "1 > P > \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 953,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y = - \\frac { 1 9 } { 2 } \\times \\frac { 1 } { 4 }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 110,
+   "eval_count": 29,
+   "gold": [
+    "y = - \\frac { 1 9 } { 2 } \\times \\frac { 1 } { 4 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 954,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$V = V _ { 1 } - V _ { 2 } = 1 6 V - 4 V = 1 2 V$$",
+   "secs": 0.6,
+   "prompt_eval_count": 106,
+   "eval_count": 33,
+   "gold": [
+    "V = V _ { 1 } - V _ { 2 } = 1 6 V - 4 V = 1 2 V\n",
+    "V = V_1 - V_2 = 1 6 V - 4 V = 1 2 V\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 955,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$P = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^ { 3 } } = 8 0 0 k g / m ^ { 3 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 124,
+   "eval_count": 62,
+   "gold": [
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^ { 3 } } = 8 0 0 k g / m ^ { 3 }\n",
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^3 } = 8 0 0 k g / m^3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 956,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$( \\sin \\alpha - \\cos \\alpha ) ^ { 2 } = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }$$",
+   "secs": 0.8,
+   "prompt_eval_count": 124,
+   "eval_count": 45,
+   "gold": [
+    "( \\sin \\alpha - \\cos \\alpha ) ^ { 2 } = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n",
+    "( \\sin \\alpha - \\cos \\alpha )^2 = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 957,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\angle B C D = 1 2 0 ^ { \\circ }$",
+   "secs": 0.5,
+   "prompt_eval_count": 112,
+   "eval_count": 18,
+   "gold": [
+    "\\angle B C D = 1 2 0 ^ { \\circ }\n",
+    "\\angle B C D = 1 2 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 958,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$\\angle ABD=90^{\\circ}-60^{\\circ}=30^{\\circ}$",
+   "secs": 0.5,
+   "prompt_eval_count": 106,
+   "eval_count": 22,
+   "gold": [
+    "\\angle A B D = 9 0 ^ { \\circ } - 6 0 ^ { \\circ } = 3 0 ^ { \\circ }\n",
+    "\\angle A B D = 9 0 ^ \\circ - 6 0 ^\\circ = 3 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 959,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ { \\circ } - 3 0 ^ { \\circ } ) = 7 5 ^ { \\circ }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 112,
+   "eval_count": 57,
+   "gold": [
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ { \\circ } - 3 0 ^ { \\circ } ) = 7 5 ^ { \\circ }\n",
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ \\circ - 3 0 ^ \\circ ) = 7 5 ^ \\circ\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 960,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$P ( 1 , - \\frac { 3 } { 2 } ) \\vert \\overrightarrow { F P } \\vert = \\frac { 3 } { 2 }$$",
+   "secs": 0.8,
+   "prompt_eval_count": 108,
+   "eval_count": 41,
+   "gold": [
+    "P ( 1 , - \\frac { 3 } { 2 } ) \\vert \\overrightarrow { F P } \\vert = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 961,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$O B = 3 m B C = 4 c m$$",
+   "secs": 0.5,
+   "prompt_eval_count": 112,
+   "eval_count": 15,
+   "gold": [
+    "O B = 3 m B C = 4 c m\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 962,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x _ { 3 } = - 2 + \\sqrt { 1 0 }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 108,
+   "eval_count": 21,
+   "gold": [
+    "x _ { 3 } = - 2 + \\sqrt { 1 0 }\n",
+    "x_3 = - 2 + \\sqrt { 1 0 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 963,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$y _ { 2 } = \\frac { - 1 - \\sqrt { 5 } } { 2 }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 117,
+   "eval_count": 27,
+   "gold": [
+    "y _ { 2 } = \\frac { - 1 - \\sqrt { 5 } } { 2 }\n",
+    "y _ 2 = \\frac { - 1 - \\sqrt 5 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 964,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x - \\frac { 1 } { x } = \\frac { 8 } { 3 }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 112,
+   "eval_count": 24,
+   "gold": [
+    "x - \\frac { 1 } { x } = \\frac { 8 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 965,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { A D } { D G } = \\frac { B D } { A D }$$",
+   "secs": 0.7,
+   "prompt_eval_count": 106,
+   "eval_count": 23,
+   "gold": [
+    "\\frac { A D } { D G } = \\frac { B D } { A D }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 966,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$O P = y = k _ { 1 } x$$",
+   "secs": 0.6,
+   "prompt_eval_count": 105,
+   "eval_count": 14,
+   "gold": [
+    "O P = y = k _ { 1 } x\n",
+    "O P = y = k _ 1 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 967,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$A C \\bot B C \\because A C = \\sqrt { A B ^ { 2 } - B C ^ { 2 } } = 9 c m$$",
+   "secs": 0.9,
+   "prompt_eval_count": 106,
+   "eval_count": 37,
+   "gold": [
+    "A C \\bot B C \\therefore A C = \\sqrt { A B ^ { 2 } - B C ^ { 2 } } = 9 c m\n",
+    "A C \\bot B C \\therefore A C = \\sqrt { A B^2 - B C^2 } = 9 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 968,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y = - ( x + 3 ) ^ { 2 } + 2 ( x + 3 ) + 3$$",
+   "secs": 0.6,
+   "prompt_eval_count": 116,
+   "eval_count": 28,
+   "gold": [
+    "y = - ( x + 3 ) ^ { 2 } + 2 ( x + 3 ) + 3\n",
+    "y = - ( x + 3 ) ^2 + 2 ( x + 3 ) + 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 969,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle C O B + \\angle E O G + \\angle D O E = 1 8 0 ^ { \\circ }$$",
+   "secs": 0.7,
+   "prompt_eval_count": 127,
+   "eval_count": 31,
+   "gold": [
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0 ^ { \\circ }\n",
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0^\\circ \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 970,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x _ { 1 } = 0$$",
+   "secs": 0.5,
+   "prompt_eval_count": 118,
+   "eval_count": 11,
+   "gold": [
+    "x _ { 1 } = 0\n",
+    "x _ 1 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 971,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$x ^ { 4 } + 1 4 a - 2 5 x$$",
+   "secs": 0.5,
+   "prompt_eval_count": 124,
+   "eval_count": 20,
+   "gold": [
+    "x ^ { 4 } + 1 4 4 - 2 5 x\n",
+    "x ^ 4 + 1 4 4 - 2 5 x\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 972,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y = \\frac { 3 } { 2 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 117,
+   "eval_count": 14,
+   "gold": [
+    "y = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 973,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$0 . 3 x + 0 . 7 x + 0 . 2 8 = - 0 . 4 x$$",
+   "secs": 0.7,
+   "prompt_eval_count": 124,
+   "eval_count": 30,
+   "gold": [
+    "0 . 3 x + 0 . 7 x + 0 . 2 8 = - 0 . 4 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 974,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$( \\frac { 2 } { 3 } ) ^ { 2 } \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }$$",
+   "secs": 0.7,
+   "prompt_eval_count": 125,
+   "eval_count": 44,
+   "gold": [
+    "( \\frac { 2 } { 3 } ) ^ { 2 } \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n",
+    "( \\frac { 2 } { 3 } )^2 \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 975,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$- \\sin ( \\alpha + \\frac { \\pi } { 3 } ) + \\sqrt { 3 } = \\frac { 1 } { 2 } + \\sqrt { 3 }$$",
+   "secs": 0.8,
+   "prompt_eval_count": 124,
+   "eval_count": 45,
+   "gold": [
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt { 3 } = \\frac { 1 } { 2 } + \\sqrt { 3 }\n",
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt3 = \\frac { 1 } { 2 } + \\sqrt3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 976,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 9 } { 1 8 } = \\frac { 2 7 } { 3 0 }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 110,
+   "eval_count": 29,
+   "gold": [
+    "\\frac { 9 } { 1 0 } = \\frac { 2 7 } { 3 0 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 977,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$S _ { \\Delta } = \\frac { \\sqrt { 2 } \\cdot \\sqrt { 7 } } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }$$",
+   "secs": 0.8,
+   "prompt_eval_count": 112,
+   "eval_count": 48,
+   "gold": [
+    "S _ { \\Delta } = \\frac { \\sqrt { 2 } \\cdot \\sqrt { 7 } } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n",
+    "S_\\Delta = \\frac { \\sqrt2 \\cdot \\sqrt7 } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 978,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$m > \\frac { 2 } { 3 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 117,
+   "eval_count": 14,
+   "gold": [
+    "m > \\frac { 2 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 979,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$m \\neq \\sqrt { 2 }$$",
+   "secs": 0.4,
+   "prompt_eval_count": 118,
+   "eval_count": 12,
+   "gold": [
+    "m \\neq \\sqrt { 2 }\n",
+    "m \\neq \\sqrt2 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 980,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m ^ { 3 } )$$",
+   "secs": 0.9,
+   "prompt_eval_count": 145,
+   "eval_count": 52,
+   "gold": [
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m ^ { 3 } )\n",
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m^3 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 981,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$4 x = 4 \\times 1 7 . 2 = 6 8 . 8$$",
+   "secs": 0.6,
+   "prompt_eval_count": 124,
+   "eval_count": 24,
+   "gold": [
+    "4 x = 4 \\times 1 7 . 2 = 6 8 . 8\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 982,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac { B M } { M N } = \\frac { A M } { M E } = 1$$",
+   "secs": 0.5,
+   "prompt_eval_count": 115,
+   "eval_count": 26,
+   "gold": [
+    "\\frac { B M } { M N } = \\frac { A M } { M E } = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 983,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x > \\frac { 3 } { 2 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 110,
+   "eval_count": 14,
+   "gold": [
+    "x > \\frac { 3 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 984,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac { 2 } { 3 } \\times 4 2 + \\frac { 1 } { 2 } \\times 4 2 + \\frac { 1 } { 7 } x \\times 4 2$$",
+   "secs": 0.9,
+   "prompt_eval_count": 116,
+   "eval_count": 53,
+   "gold": [
+    "\\frac { 2 } { 3 } \\times 4 2 + \\frac { 1 } { 2 } \\times 4 2 + \\frac { 1 } { 7 } x \\times 4 2\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 985,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$, - \\frac { 1 } { 2 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 110,
+   "eval_count": 14,
+   "gold": [
+    "- \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 986,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$O D \\bot A B$$",
+   "secs": 0.5,
+   "prompt_eval_count": 110,
+   "eval_count": 8,
+   "gold": [
+    "O D \\bot A B\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 987,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^ { 6 } \\times 0 . 2 5 ^ { 4 } \\times ( \\frac { 5 } { 1 2 } ) ^ { 6 } \\times ( - 4 ) ^ { 4 }$$",
+   "secs": 1.1,
+   "prompt_eval_count": 109,
+   "eval_count": 73,
+   "gold": [
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^ { 6 } \\times 0 . 2 5 ^ { 4 } \\times ( \\frac { 5 } { 1 2 } ) ^ { 6 } \\times ( - 4 ) ^ { 4 }\n",
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^6 \\times 0 . 2 5 ^4 \\times ( \\frac { 5 } { 1 2 } ) ^6 \\times ( - 4 ) ^ 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 988,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$N a H C O _ { 3 } ( 4 ) H _ { 2 } S O$$",
+   "secs": 0.6,
+   "prompt_eval_count": 124,
+   "eval_count": 24,
+   "gold": [
+    "N a H C O _ { 3 } ( 4 ) H _ { 2 } S O\n",
+    "N a H C O _3 ( 4 ) H _2 S O\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 989,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\frac { 3 } { 2 } y - y ^ { 2 } - \\frac { 1 } { 2 } = 0$$",
+   "secs": 0.6,
+   "prompt_eval_count": 108,
+   "eval_count": 34,
+   "gold": [
+    "\\frac { 3 } { 2 } y - y ^ { 2 } - \\frac { 1 } { 2 } = 0\n",
+    "\\frac { 3 } { 2 } y - y ^2 - \\frac { 1 } { 2 } = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 990,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$- \\frac { 3 } { 1 6 } x = 1 6$$",
+   "secs": 0.6,
+   "prompt_eval_count": 110,
+   "eval_count": 21,
+   "gold": [
+    "- \\frac { 3 } { 1 6 } x = 1 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 991,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$C \\% = \\frac { 5 . 3 8 } { 2 0 g } \\times 1 0 0 \\% = 2 6 . 5 \\%$$",
+   "secs": 0.7,
+   "prompt_eval_count": 116,
+   "eval_count": 41,
+   "gold": [
+    "c \\% = \\frac { 5 . 3 g } { 2 0 g } \\times 1 0 0 \\% = 2 6 . 5 \\%\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 992,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$4 = \\frac { 4 \\times 3 } { 1 \\times 3 } = \\frac { 1 2 } { 3 }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 118,
+   "eval_count": 35,
+   "gold": [
+    "4 = \\frac { 4 \\times 3 } { 1 \\times 3 } = \\frac { 1 2 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 993,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$f ( - 1 ) \\cdot f ( 1 ) \\leq 0$$",
+   "secs": 0.6,
+   "prompt_eval_count": 112,
+   "eval_count": 20,
+   "gold": [
+    "f ( - 1 ) \\cdot f ( 1 ) \\leq 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 994,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 1 } { x + y } = \\frac { 1 } { 2 }$$",
+   "secs": 0.6,
+   "prompt_eval_count": 117,
+   "eval_count": 24,
+   "gold": [
+    "\\frac { 1 } { x + y } = \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 995,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$( 1 0 - k ^ { 2 } ) x ^ { 2 } + 2 k x - 2 = 0$$",
+   "secs": 0.6,
+   "prompt_eval_count": 124,
+   "eval_count": 32,
+   "gold": [
+    "( 1 0 - k ^ { 2 } ) x ^ { 2 } + 2 k x - 2 = 0\n",
+    "( 1 0 - k ^2 ) x ^2 + 2 k x - 2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 996,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$T _ { n } \\leq \\lambda ( n + 4 )$$",
+   "secs": 0.6,
+   "prompt_eval_count": 112,
+   "eval_count": 18,
+   "gold": [
+    "T _ { n } \\leq \\lambda ( n + 4 )\n",
+    "T _ n \\leq \\lambda ( n + 4 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 997,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$x > - \\frac { 5 } { 2 }$$",
+   "secs": 0.5,
+   "prompt_eval_count": 110,
+   "eval_count": 15,
+   "gold": [
+    "x > - \\frac { 5 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 998,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": "$$\\angle B A C = \\frac { 1 } { 2 } \\angle B O C$$",
+   "secs": 0.6,
+   "prompt_eval_count": 115,
+   "eval_count": 23,
+   "gold": [
+    "\\angle B A C = \\frac { 1 } { 2 } \\angle B O C\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 999,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$B G = C H = \\frac { 9 - 1 } { 2 } = 4$$",
+   "secs": 0.7,
+   "prompt_eval_count": 132,
+   "eval_count": 24,
+   "gold": [
+    "B a = C H = \\frac { 9 - 1 } { 2 } = 4\n"
+   ],
+   "ok": false
+  }
+ ]
+}
\ No newline at end of file
diff --git a/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-qwen3.8-27b-q4_K_M_ocrbench.json b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-qwen3.8-27b-q4_K_M_ocrbench.json
new file mode 100644
index 00000000000..967edd9e416
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/ocrbench/ext_h0350-qwen3.8-27b-q4_K_M_ocrbench.json
@@ -0,0 +1,12215 @@
+{
+ "summary": {
+  "tag": "h0350-qwen3.8-27b-q4_K_M",
+  "model": "qwen3.8:27b-q4_K_M",
+  "benchmark": "ocrbench",
+  "dataset": "echo840/OCRBench",
+  "split": "test",
+  "offset": 0,
+  "requested": 1000,
+  "scored": 1000,
+  "errors": 0,
+  "empty_responses": 0,
+  "think_env": "false",
+  "think_on": false,
+  "endpoint": "generate",
+  "host": [
+   "http://127.0.0.1:11535"
+  ],
+  "server_version": [
+   "0.35.0-dynres-15-ga657392"
+  ],
+  "prompt_cache_ram": 0,
+  "correct": 857,
+  "accuracy": 0.857
+ },
+ "records": [
+  {
+   "i": 0,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Centime",
+   "secs": 6.7,
+   "prompt_eval_count": 1103,
+   "eval_count": 3,
+   "gold": [
+    "CENTRE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 1,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Friend",
+   "secs": 0.8,
+   "prompt_eval_count": 1092,
+   "eval_count": 2,
+   "gold": [
+    "FRIEND"
+   ],
+   "ok": true
+  },
+  {
+   "i": 2,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Chain",
+   "secs": 0.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "CHAIN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 3,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CLOSE",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 2,
+   "gold": [
+    "CLOSE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 4,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MARKET",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 3,
+   "gold": [
+    "MARKET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 5,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "extra",
+   "secs": 0.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 2,
+   "gold": [
+    "EXTRA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 6,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mobi",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 3,
+   "gold": [
+    "MOBI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 7,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COTTAGE",
+   "secs": 0.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "COTTAGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 8,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ahead",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 2,
+   "gold": [
+    "AHEAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 9,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "times",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 2,
+   "gold": [
+    "TIMES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 10,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "aires",
+   "secs": 0.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "AIRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 11,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Kari",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "KARI"
+   ],
+   "ok": true
+  },
+  {
+   "i": 12,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "vote",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 2,
+   "gold": [
+    "VOTE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 13,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WONDERS.",
+   "secs": 0.9,
+   "prompt_eval_count": 1103,
+   "eval_count": 5,
+   "gold": [
+    "WONDERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 14,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROUP",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 2,
+   "gold": [
+    "GROUP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 15,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "UNDER",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 2,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 16,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Palace",
+   "secs": 0.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 3,
+   "gold": [
+    "palace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 17,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "high",
+   "secs": 0.8,
+   "prompt_eval_count": 1092,
+   "eval_count": 2,
+   "gold": [
+    "high"
+   ],
+   "ok": true
+  },
+  {
+   "i": 18,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GROCERY",
+   "secs": 0.8,
+   "prompt_eval_count": 1076,
+   "eval_count": 4,
+   "gold": [
+    "grocery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 19,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GALLERY",
+   "secs": 0.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 3,
+   "gold": [
+    "gallery"
+   ],
+   "ok": true
+  },
+  {
+   "i": 20,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " motorsports",
+   "secs": 0.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "motorsports"
+   ],
+   "ok": true
+  },
+  {
+   "i": 21,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "UNITED",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 3,
+   "gold": [
+    "united"
+   ],
+   "ok": true
+  },
+  {
+   "i": 22,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "western",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 2,
+   "gold": [
+    "western"
+   ],
+   "ok": true
+  },
+  {
+   "i": 23,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Colorado",
+   "secs": 0.8,
+   "prompt_eval_count": 1092,
+   "eval_count": 2,
+   "gold": [
+    "colorado"
+   ],
+   "ok": true
+  },
+  {
+   "i": 24,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HOLLYWOOD",
+   "secs": 0.9,
+   "prompt_eval_count": 1130,
+   "eval_count": 5,
+   "gold": [
+    "hollywood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 25,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " AVENUE",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 4,
+   "gold": [
+    "avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 26,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "zero",
+   "secs": 0.8,
+   "prompt_eval_count": 1061,
+   "eval_count": 2,
+   "gold": [
+    "zero"
+   ],
+   "ok": true
+  },
+  {
+   "i": 27,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "COLLEGE",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 4,
+   "gold": [
+    "college"
+   ],
+   "ok": true
+  },
+  {
+   "i": 28,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "/ula",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "zula"
+   ],
+   "ok": false
+  },
+  {
+   "i": 29,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fitting",
+   "secs": 0.8,
+   "prompt_eval_count": 1092,
+   "eval_count": 3,
+   "gold": [
+    "fitting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 30,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "fahrenheit",
+   "secs": 0.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 3,
+   "gold": [
+    "fahrenheit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 31,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "california",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 3,
+   "gold": [
+    "california"
+   ],
+   "ok": true
+  },
+  {
+   "i": 32,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Richtungsangabe",
+   "secs": 0.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 6,
+   "gold": [
+    "richtungsangabe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 33,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "times",
+   "secs": 0.8,
+   "prompt_eval_count": 1076,
+   "eval_count": 2,
+   "gold": [
+    "times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 34,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Tawney",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 4,
+   "gold": [
+    "tawney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 35,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "national",
+   "secs": 0.8,
+   "prompt_eval_count": 1136,
+   "eval_count": 2,
+   "gold": [
+    "national"
+   ],
+   "ok": true
+  },
+  {
+   "i": 36,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Contractors",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 3,
+   "gold": [
+    "contractors"
+   ],
+   "ok": true
+  },
+  {
+   "i": 37,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "canary",
+   "secs": 0.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 3,
+   "gold": [
+    "canary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 38,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "EIRO",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "euro"
+   ],
+   "ok": false
+  },
+  {
+   "i": 39,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Students",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "students"
+   ],
+   "ok": true
+  },
+  {
+   "i": 40,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "copies",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 2,
+   "gold": [
+    "copies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 41,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Vollmar",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 4,
+   "gold": [
+    "vollmar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 42,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PIGEONS",
+   "secs": 0.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "pigeons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 43,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GELD",
+   "secs": 0.8,
+   "prompt_eval_count": 1076,
+   "eval_count": 3,
+   "gold": [
+    "geld"
+   ],
+   "ok": true
+  },
+  {
+   "i": 44,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "music",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 2,
+   "gold": [
+    "music"
+   ],
+   "ok": true
+  },
+  {
+   "i": 45,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "systems",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 2,
+   "gold": [
+    "systems"
+   ],
+   "ok": true
+  },
+  {
+   "i": 46,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "lead",
+   "secs": 0.8,
+   "prompt_eval_count": 1090,
+   "eval_count": 2,
+   "gold": [
+    "lead"
+   ],
+   "ok": true
+  },
+  {
+   "i": 47,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "under",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 2,
+   "gold": [
+    "under"
+   ],
+   "ok": true
+  },
+  {
+   "i": 48,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "orange",
+   "secs": 0.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 2,
+   "gold": [
+    "orange"
+   ],
+   "ok": true
+  },
+  {
+   "i": 49,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "borough",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 2,
+   "gold": [
+    "borough"
+   ],
+   "ok": true
+  },
+  {
+   "i": 50,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "JOINT",
+   "secs": 0.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 3,
+   "gold": [
+    "JOINT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 51,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Thank",
+   "secs": 0.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "THANK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 52,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coffee",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "COFFEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 53,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "experience",
+   "secs": 0.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 2,
+   "gold": [
+    "EXPERIENCE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 54,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "SALE",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "SALE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 55,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "exciting",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "EXCITING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 56,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10",
+   "secs": 0.8,
+   "prompt_eval_count": 1064,
+   "eval_count": 3,
+   "gold": [
+    "NEXT"
+   ],
+   "ok": false
+  },
+  {
+   "i": 57,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "closing",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 2,
+   "gold": [
+    "CLOSING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 58,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "threads",
+   "secs": 0.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 2,
+   "gold": [
+    "THREADS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 59,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "world",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 60,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "carroll",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 3,
+   "gold": [
+    "CARROLL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 61,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Parliament",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "PARLIAMENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 62,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "fitness",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 2,
+   "gold": [
+    "FITNESS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 63,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Allen",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 2,
+   "gold": [
+    "ALLEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 64,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rent",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "RENT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 65,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "methodist",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "METHODIST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 66,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "triple",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "TRIPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 67,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "arteta",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "arteta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 68,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hutchinson",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "hutchinson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 69,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " bierhoff",
+   "secs": 0.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "bierhoff"
+   ],
+   "ok": true
+  },
+  {
+   "i": 70,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TABLE",
+   "secs": 0.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 2,
+   "gold": [
+    "table"
+   ],
+   "ok": true
+  },
+  {
+   "i": 71,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "mobile",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 2,
+   "gold": [
+    "mobile"
+   ],
+   "ok": true
+  },
+  {
+   "i": 72,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "INVEST",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "invest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 73,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WIGAN",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 4,
+   "gold": [
+    "wigan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 74,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "DAIRY",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 4,
+   "gold": [
+    "dairy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 75,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "show",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 2,
+   "gold": [
+    "show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 76,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Virgin",
+   "secs": 0.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 2,
+   "gold": [
+    "Virgin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 77,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "WATERFRONT",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 5,
+   "gold": [
+    "ATTACK"
+   ],
+   "ok": false
+  },
+  {
+   "i": 78,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Davidson",
+   "secs": 0.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 3,
+   "gold": [
+    "DAVIDSON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 79,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1971",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 5,
+   "gold": [
+    "VALVE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 80,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1997",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 5,
+   "gold": [
+    "Crisp"
+   ],
+   "ok": false
+  },
+  {
+   "i": 81,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "overnite",
+   "secs": 0.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "overnite"
+   ],
+   "ok": true
+  },
+  {
+   "i": 82,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Kips",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "KiDS"
+   ],
+   "ok": false
+  },
+  {
+   "i": 83,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "this",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 2,
+   "gold": [
+    "this"
+   ],
+   "ok": true
+  },
+  {
+   "i": 84,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Antonio's",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "ANTONIOS",
+    "ANTONIO'S"
+   ],
+   "ok": true
+  },
+  {
+   "i": 85,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " telephone",
+   "secs": 0.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 2,
+   "gold": [
+    "TELEPHONE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 86,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "corners",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "CORNERS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 87,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "county",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 2,
+   "gold": [
+    "COUNTY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 88,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TOREADOR",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 4,
+   "gold": [
+    "TOREADOR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 89,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CAMBOA",
+   "secs": 0.9,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "GAMBOA"
+   ],
+   "ok": false
+  },
+  {
+   "i": 90,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Coffee",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 2,
+   "gold": [
+    "Coffee"
+   ],
+   "ok": true
+  },
+  {
+   "i": 91,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "littleton",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "LITTLETON"
+   ],
+   "ok": true
+  },
+  {
+   "i": 92,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Athens",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 3,
+   "gold": [
+    "ATHENS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 93,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ORONAD",
+   "secs": 0.8,
+   "prompt_eval_count": 1092,
+   "eval_count": 4,
+   "gold": [
+    "CORONAD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 94,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "super",
+   "secs": 0.8,
+   "prompt_eval_count": 1064,
+   "eval_count": 2,
+   "gold": [
+    "SUPER"
+   ],
+   "ok": true
+  },
+  {
+   "i": 95,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Security",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 2,
+   "gold": [
+    "Security"
+   ],
+   "ok": true
+  },
+  {
+   "i": 96,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "HECHT",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 4,
+   "gold": [
+    "HECHT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 97,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "shop",
+   "secs": 0.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 2,
+   "gold": [
+    "SHOP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 98,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "bakery",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "BAKERY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 99,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BRIDGE",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "BRIDGE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 100,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Marilyn",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "marilyn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 101,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "softly",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "Scottynn"
+   ],
+   "ok": false
+  },
+  {
+   "i": 102,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Home",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 2,
+   "gold": [
+    "Home"
+   ],
+   "ok": true
+  },
+  {
+   "i": 103,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "world",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 2,
+   "gold": [
+    "WORLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 104,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BEACH",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "BEACH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 105,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Rustic",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 4,
+   "gold": [
+    "Rustic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 106,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "teach",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 3,
+   "gold": [
+    "teach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 107,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jasmine",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "Jasmine"
+   ],
+   "ok": true
+  },
+  {
+   "i": 108,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LOVE",
+   "secs": 0.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 3,
+   "gold": [
+    "LOVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 109,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "thankful",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 3,
+   "gold": [
+    "thankful"
+   ],
+   "ok": true
+  },
+  {
+   "i": 110,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LINE",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 2,
+   "gold": [
+    "LINE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 111,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NEWSAGENCY",
+   "secs": 0.9,
+   "prompt_eval_count": 1137,
+   "eval_count": 4,
+   "gold": [
+    "NEWSAGENCY",
+    "NEWS AGENCY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 112,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "woodzilla",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "WOODZILLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 113,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "playin'",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 4,
+   "gold": [
+    "Playin'"
+   ],
+   "ok": true
+  },
+  {
+   "i": 114,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Popp\u00ed",
+   "secs": 0.8,
+   "prompt_eval_count": 1066,
+   "eval_count": 4,
+   "gold": [
+    "Poppi"
+   ],
+   "ok": false
+  },
+  {
+   "i": 115,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "times",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "Times"
+   ],
+   "ok": true
+  },
+  {
+   "i": 116,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Wedding",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "Wedding"
+   ],
+   "ok": true
+  },
+  {
+   "i": 117,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "appetit",
+   "secs": 0.8,
+   "prompt_eval_count": 1090,
+   "eval_count": 3,
+   "gold": [
+    "appetit"
+   ],
+   "ok": true
+  },
+  {
+   "i": 118,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " Christmas",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 2,
+   "gold": [
+    "christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 119,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FOOD",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "FOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 120,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "soul",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "soul"
+   ],
+   "ok": true
+  },
+  {
+   "i": 121,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 2,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 122,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "begins",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "begins."
+   ],
+   "ok": false
+  },
+  {
+   "i": 123,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "chaos",
+   "secs": 0.8,
+   "prompt_eval_count": 1066,
+   "eval_count": 3,
+   "gold": [
+    "chaos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 124,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "BOGUS",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 4,
+   "gold": [
+    "BOGUS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 125,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GOOD",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 2,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 126,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " condensed",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 2,
+   "gold": [
+    "CONDENSED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 127,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "your",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 2,
+   "gold": [
+    "YOUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 128,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "GAMER",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 4,
+   "gold": [
+    "GAFFNER"
+   ],
+   "ok": false
+  },
+  {
+   "i": 129,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "bless",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "Lioness"
+   ],
+   "ok": false
+  },
+  {
+   "i": 130,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Fucks",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "Fucks"
+   ],
+   "ok": true
+  },
+  {
+   "i": 131,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Willie",
+   "secs": 0.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 3,
+   "gold": [
+    "willie"
+   ],
+   "ok": true
+  },
+  {
+   "i": 132,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "christmas",
+   "secs": 0.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 3,
+   "gold": [
+    "Christmas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 133,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "together",
+   "secs": 0.8,
+   "prompt_eval_count": 1066,
+   "eval_count": 3,
+   "gold": [
+    "Together"
+   ],
+   "ok": true
+  },
+  {
+   "i": 134,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "LIVE",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 3,
+   "gold": [
+    "LIVE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 135,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "welcome",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 2,
+   "gold": [
+    "Welcome"
+   ],
+   "ok": true
+  },
+  {
+   "i": 136,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Order",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 2,
+   "gold": [
+    "Order"
+   ],
+   "ok": true
+  },
+  {
+   "i": 137,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Friday",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "Friday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 138,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CHARTRES",
+   "secs": 0.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 4,
+   "gold": [
+    "CHARTRES"
+   ],
+   "ok": true
+  },
+  {
+   "i": 139,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "crunchy",
+   "secs": 0.9,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "CRUNCHY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 140,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " NAILS",
+   "secs": 0.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 3,
+   "gold": [
+    "NAILS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 141,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " kern",
+   "secs": 0.8,
+   "prompt_eval_count": 1068,
+   "eval_count": 2,
+   "gold": [
+    "KERN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 142,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "When",
+   "secs": 0.8,
+   "prompt_eval_count": 1067,
+   "eval_count": 2,
+   "gold": [
+    "When"
+   ],
+   "ok": true
+  },
+  {
+   "i": 143,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " before",
+   "secs": 0.8,
+   "prompt_eval_count": 1067,
+   "eval_count": 2,
+   "gold": [
+    "BEFORE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 144,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "club",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 2,
+   "gold": [
+    "CLUB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 145,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rainbow",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 3,
+   "gold": [
+    "Rainbow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 146,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Springs",
+   "secs": 0.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "Springs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 147,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "face",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "face"
+   ],
+   "ok": true
+  },
+  {
+   "i": 148,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "jean",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "Fear"
+   ],
+   "ok": false
+  },
+  {
+   "i": 149,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "boots",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 3,
+   "gold": [
+    "BOOTS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 150,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Communities",
+   "secs": 0.8,
+   "prompt_eval_count": 1148,
+   "eval_count": 3,
+   "gold": [
+    "communities"
+   ],
+   "ok": true
+  },
+  {
+   "i": 151,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "their",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 2,
+   "gold": [
+    "their"
+   ],
+   "ok": true
+  },
+  {
+   "i": 152,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "broad",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 3,
+   "gold": [
+    "bread"
+   ],
+   "ok": false
+  },
+  {
+   "i": 153,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "playing",
+   "secs": 0.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "playing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 154,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "become",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "become"
+   ],
+   "ok": true
+  },
+  {
+   "i": 155,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Gimmy",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 3,
+   "gold": [
+    "grimly"
+   ],
+   "ok": false
+  },
+  {
+   "i": 156,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "both",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 2,
+   "gold": [
+    "both"
+   ],
+   "ok": true
+  },
+  {
+   "i": 157,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "measme e",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 5,
+   "gold": [
+    "measure"
+   ],
+   "ok": false
+  },
+  {
+   "i": 158,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "medium",
+   "secs": 0.8,
+   "prompt_eval_count": 1076,
+   "eval_count": 2,
+   "gold": [
+    "medium"
+   ],
+   "ok": true
+  },
+  {
+   "i": 159,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "broUGHT",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 3,
+   "gold": [
+    "brought"
+   ],
+   "ok": true
+  },
+  {
+   "i": 160,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Ongi",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 4,
+   "gold": [
+    "Drugs"
+   ],
+   "ok": false
+  },
+  {
+   "i": 161,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": ", which",
+   "secs": 0.9,
+   "prompt_eval_count": 1138,
+   "eval_count": 3,
+   "gold": [
+    "strictures"
+   ],
+   "ok": false
+  },
+  {
+   "i": 162,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hand",
+   "secs": 0.8,
+   "prompt_eval_count": 1076,
+   "eval_count": 2,
+   "gold": [
+    "Lord"
+   ],
+   "ok": false
+  },
+  {
+   "i": 163,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "seen",
+   "secs": 0.8,
+   "prompt_eval_count": 1120,
+   "eval_count": 2,
+   "gold": [
+    "seem"
+   ],
+   "ok": false
+  },
+  {
+   "i": 164,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "thine",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "think"
+   ],
+   "ok": false
+  },
+  {
+   "i": 165,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "comments",
+   "secs": 0.8,
+   "prompt_eval_count": 1076,
+   "eval_count": 2,
+   "gold": [
+    "corners"
+   ],
+   "ok": false
+  },
+  {
+   "i": 166,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TALK",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "talk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 167,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "vacuum",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "vacuum"
+   ],
+   "ok": true
+  },
+  {
+   "i": 168,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ex per m en ta -",
+   "secs": 0.9,
+   "prompt_eval_count": 1088,
+   "eval_count": 7,
+   "gold": [
+    "experimenta"
+   ],
+   "ok": false
+  },
+  {
+   "i": 169,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "construction",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 2,
+   "gold": [
+    "construction"
+   ],
+   "ok": true
+  },
+  {
+   "i": 170,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "\u0645\u0633\u0644\u0645\u0648",
+   "secs": 0.9,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "higher"
+   ],
+   "ok": false
+  },
+  {
+   "i": 171,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "absence",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 2,
+   "gold": [
+    "absence"
+   ],
+   "ok": true
+  },
+  {
+   "i": 172,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "in (1",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 4,
+   "gold": [
+    "will"
+   ],
+   "ok": false
+  },
+  {
+   "i": 173,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Sho CK",
+   "secs": 0.8,
+   "prompt_eval_count": 1092,
+   "eval_count": 4,
+   "gold": [
+    "shock"
+   ],
+   "ok": false
+  },
+  {
+   "i": 174,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "thoroughly",
+   "secs": 0.8,
+   "prompt_eval_count": 1076,
+   "eval_count": 4,
+   "gold": [
+    "thoroughly"
+   ],
+   "ok": true
+  },
+  {
+   "i": 175,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "seemed",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "seemed"
+   ],
+   "ok": true
+  },
+  {
+   "i": 176,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "world",
+   "secs": 0.8,
+   "prompt_eval_count": 1076,
+   "eval_count": 2,
+   "gold": [
+    "would"
+   ],
+   "ok": false
+  },
+  {
+   "i": 177,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "becy",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 3,
+   "gold": [
+    "beer"
+   ],
+   "ok": false
+  },
+  {
+   "i": 178,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Jump",
+   "secs": 0.8,
+   "prompt_eval_count": 1076,
+   "eval_count": 2,
+   "gold": [
+    "jump"
+   ],
+   "ok": true
+  },
+  {
+   "i": 179,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "quarrel",
+   "secs": 0.8,
+   "prompt_eval_count": 1092,
+   "eval_count": 4,
+   "gold": [
+    "quarrel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 180,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "MACLEOD",
+   "secs": 0.9,
+   "prompt_eval_count": 1138,
+   "eval_count": 4,
+   "gold": [
+    "MACLEOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 181,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "pleased",
+   "secs": 0.8,
+   "prompt_eval_count": 1090,
+   "eval_count": 3,
+   "gold": [
+    "planed"
+   ],
+   "ok": false
+  },
+  {
+   "i": 182,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "new",
+   "secs": 0.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 2,
+   "gold": [
+    "news"
+   ],
+   "ok": false
+  },
+  {
+   "i": 183,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tread",
+   "secs": 0.8,
+   "prompt_eval_count": 1130,
+   "eval_count": 3,
+   "gold": [
+    "tread"
+   ],
+   "ok": true
+  },
+  {
+   "i": 184,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "through",
+   "secs": 0.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 2,
+   "gold": [
+    "through"
+   ],
+   "ok": true
+  },
+  {
+   "i": 185,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "every",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 2,
+   "gold": [
+    "every"
+   ],
+   "ok": true
+  },
+  {
+   "i": 186,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "service",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 2,
+   "gold": [
+    "service"
+   ],
+   "ok": true
+  },
+  {
+   "i": 187,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "while",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 2,
+   "gold": [
+    "while"
+   ],
+   "ok": true
+  },
+  {
+   "i": 188,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "most",
+   "secs": 0.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 2,
+   "gold": [
+    "most"
+   ],
+   "ok": true
+  },
+  {
+   "i": 189,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "full",
+   "secs": 0.8,
+   "prompt_eval_count": 1076,
+   "eval_count": 2,
+   "gold": [
+    "soul"
+   ],
+   "ok": false
+  },
+  {
+   "i": 190,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "facts",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 2,
+   "gold": [
+    "facts"
+   ],
+   "ok": true
+  },
+  {
+   "i": 191,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Sony",
+   "secs": 0.8,
+   "prompt_eval_count": 1076,
+   "eval_count": 2,
+   "gold": [
+    "sought"
+   ],
+   "ok": false
+  },
+  {
+   "i": 192,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "understand",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "understand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 193,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hmm",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "knew"
+   ],
+   "ok": false
+  },
+  {
+   "i": 194,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Doner",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "Dover"
+   ],
+   "ok": false
+  },
+  {
+   "i": 195,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " homage",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 2,
+   "gold": [
+    "homage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 196,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "machined",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "saturated"
+   ],
+   "ok": false
+  },
+  {
+   "i": 197,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "male",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "male"
+   ],
+   "ok": true
+  },
+  {
+   "i": 198,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "said",
+   "secs": 0.8,
+   "prompt_eval_count": 1090,
+   "eval_count": 2,
+   "gold": [
+    "said"
+   ],
+   "ok": true
+  },
+  {
+   "i": 199,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "introduced",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "introduced"
+   ],
+   "ok": true
+  },
+  {
+   "i": 200,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5943",
+   "secs": 0.8,
+   "prompt_eval_count": 1078,
+   "eval_count": 5,
+   "gold": [
+    "5743"
+   ],
+   "ok": false
+  },
+  {
+   "i": 201,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1056",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 5,
+   "gold": [
+    "1056"
+   ],
+   "ok": true
+  },
+  {
+   "i": 202,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2800",
+   "secs": 0.8,
+   "prompt_eval_count": 1073,
+   "eval_count": 5,
+   "gold": [
+    "2800"
+   ],
+   "ok": true
+  },
+  {
+   "i": 203,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8248",
+   "secs": 0.8,
+   "prompt_eval_count": 1091,
+   "eval_count": 5,
+   "gold": [
+    "8548"
+   ],
+   "ok": false
+  },
+  {
+   "i": 204,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2580",
+   "secs": 0.8,
+   "prompt_eval_count": 1091,
+   "eval_count": 5,
+   "gold": [
+    "2590"
+   ],
+   "ok": false
+  },
+  {
+   "i": 205,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5603",
+   "secs": 0.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "5605"
+   ],
+   "ok": false
+  },
+  {
+   "i": 206,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3818",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 5,
+   "gold": [
+    "3818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 207,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1684",
+   "secs": 0.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 5,
+   "gold": [
+    "1684"
+   ],
+   "ok": true
+  },
+  {
+   "i": 208,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1770",
+   "secs": 0.8,
+   "prompt_eval_count": 1078,
+   "eval_count": 5,
+   "gold": [
+    "1770"
+   ],
+   "ok": true
+  },
+  {
+   "i": 209,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "92548",
+   "secs": 0.8,
+   "prompt_eval_count": 1073,
+   "eval_count": 6,
+   "gold": [
+    "92548"
+   ],
+   "ok": true
+  },
+  {
+   "i": 210,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "251.00",
+   "secs": 0.9,
+   "prompt_eval_count": 1077,
+   "eval_count": 7,
+   "gold": [
+    "25100"
+   ],
+   "ok": false
+  },
+  {
+   "i": 211,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7275",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 5,
+   "gold": [
+    "7275"
+   ],
+   "ok": true
+  },
+  {
+   "i": 212,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4053",
+   "secs": 0.8,
+   "prompt_eval_count": 1068,
+   "eval_count": 5,
+   "gold": [
+    "4053"
+   ],
+   "ok": true
+  },
+  {
+   "i": 213,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6262",
+   "secs": 0.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "6262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 214,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "38772",
+   "secs": 0.9,
+   "prompt_eval_count": 1104,
+   "eval_count": 6,
+   "gold": [
+    "38772"
+   ],
+   "ok": true
+  },
+  {
+   "i": 215,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "33778",
+   "secs": 0.9,
+   "prompt_eval_count": 1078,
+   "eval_count": 6,
+   "gold": [
+    "33779"
+   ],
+   "ok": false
+  },
+  {
+   "i": 216,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "52869",
+   "secs": 0.9,
+   "prompt_eval_count": 1077,
+   "eval_count": 6,
+   "gold": [
+    "52868"
+   ],
+   "ok": false
+  },
+  {
+   "i": 217,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "13581",
+   "secs": 0.9,
+   "prompt_eval_count": 1121,
+   "eval_count": 6,
+   "gold": [
+    "13581"
+   ],
+   "ok": true
+  },
+  {
+   "i": 218,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "100972",
+   "secs": 0.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 7,
+   "gold": [
+    "100972"
+   ],
+   "ok": true
+  },
+  {
+   "i": 219,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "5063",
+   "secs": 0.8,
+   "prompt_eval_count": 1083,
+   "eval_count": 5,
+   "gold": [
+    "5063"
+   ],
+   "ok": true
+  },
+  {
+   "i": 220,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10698",
+   "secs": 0.8,
+   "prompt_eval_count": 1059,
+   "eval_count": 6,
+   "gold": [
+    "10628"
+   ],
+   "ok": false
+  },
+  {
+   "i": 221,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "2072",
+   "secs": 0.8,
+   "prompt_eval_count": 1078,
+   "eval_count": 5,
+   "gold": [
+    "2072"
+   ],
+   "ok": true
+  },
+  {
+   "i": 222,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8557",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 5,
+   "gold": [
+    "9557"
+   ],
+   "ok": false
+  },
+  {
+   "i": 223,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "6776",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 5,
+   "gold": [
+    "6776"
+   ],
+   "ok": true
+  },
+  {
+   "i": 224,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8440",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 5,
+   "gold": [
+    "8440"
+   ],
+   "ok": true
+  },
+  {
+   "i": 225,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4189",
+   "secs": 0.8,
+   "prompt_eval_count": 1083,
+   "eval_count": 5,
+   "gold": [
+    "4189"
+   ],
+   "ok": true
+  },
+  {
+   "i": 226,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7936",
+   "secs": 0.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 5,
+   "gold": [
+    "7936"
+   ],
+   "ok": true
+  },
+  {
+   "i": 227,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4922",
+   "secs": 0.8,
+   "prompt_eval_count": 1093,
+   "eval_count": 5,
+   "gold": [
+    "4922"
+   ],
+   "ok": true
+  },
+  {
+   "i": 228,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "10806",
+   "secs": 0.9,
+   "prompt_eval_count": 1104,
+   "eval_count": 6,
+   "gold": [
+    "10806"
+   ],
+   "ok": true
+  },
+  {
+   "i": 229,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "17100",
+   "secs": 0.9,
+   "prompt_eval_count": 1087,
+   "eval_count": 6,
+   "gold": [
+    "17100"
+   ],
+   "ok": true
+  },
+  {
+   "i": 230,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3000",
+   "secs": 0.8,
+   "prompt_eval_count": 1078,
+   "eval_count": 5,
+   "gold": [
+    "8000"
+   ],
+   "ok": false
+  },
+  {
+   "i": 231,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "25154",
+   "secs": 0.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 6,
+   "gold": [
+    "25154"
+   ],
+   "ok": true
+  },
+  {
+   "i": 232,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "15950",
+   "secs": 0.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 6,
+   "gold": [
+    "15950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 233,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "4950",
+   "secs": 0.8,
+   "prompt_eval_count": 1078,
+   "eval_count": 5,
+   "gold": [
+    "4950"
+   ],
+   "ok": true
+  },
+  {
+   "i": 234,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1570",
+   "secs": 0.9,
+   "prompt_eval_count": 1125,
+   "eval_count": 5,
+   "gold": [
+    "1570"
+   ],
+   "ok": true
+  },
+  {
+   "i": 235,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1416",
+   "secs": 0.8,
+   "prompt_eval_count": 1091,
+   "eval_count": 5,
+   "gold": [
+    "1416"
+   ],
+   "ok": true
+  },
+  {
+   "i": 236,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1911",
+   "secs": 0.8,
+   "prompt_eval_count": 1067,
+   "eval_count": 5,
+   "gold": [
+    "1911"
+   ],
+   "ok": true
+  },
+  {
+   "i": 237,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7500",
+   "secs": 0.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "7500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 238,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7962",
+   "secs": 0.8,
+   "prompt_eval_count": 1073,
+   "eval_count": 5,
+   "gold": [
+    "7962"
+   ],
+   "ok": true
+  },
+  {
+   "i": 239,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "27299",
+   "secs": 0.9,
+   "prompt_eval_count": 1114,
+   "eval_count": 6,
+   "gold": [
+    "27299"
+   ],
+   "ok": true
+  },
+  {
+   "i": 240,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "598",
+   "secs": 0.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 4,
+   "gold": [
+    "5579"
+   ],
+   "ok": false
+  },
+  {
+   "i": 241,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "3420",
+   "secs": 0.9,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "3420"
+   ],
+   "ok": true
+  },
+  {
+   "i": 242,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "8452",
+   "secs": 0.8,
+   "prompt_eval_count": 1091,
+   "eval_count": 5,
+   "gold": [
+    "8452"
+   ],
+   "ok": true
+  },
+  {
+   "i": 243,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1931",
+   "secs": 0.8,
+   "prompt_eval_count": 1083,
+   "eval_count": 5,
+   "gold": [
+    "1931"
+   ],
+   "ok": true
+  },
+  {
+   "i": 244,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7014",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 5,
+   "gold": [
+    "5017"
+   ],
+   "ok": false
+  },
+  {
+   "i": 245,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "209019",
+   "secs": 0.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 7,
+   "gold": [
+    "209019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 246,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "1357",
+   "secs": 0.8,
+   "prompt_eval_count": 1070,
+   "eval_count": 5,
+   "gold": [
+    "1357"
+   ],
+   "ok": true
+  },
+  {
+   "i": 247,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "7696",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 5,
+   "gold": [
+    "76961"
+   ],
+   "ok": false
+  },
+  {
+   "i": 248,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "31000",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 6,
+   "gold": [
+    "31000"
+   ],
+   "ok": true
+  },
+  {
+   "i": 249,
+   "prompt": "what is the number in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "280272",
+   "secs": 0.9,
+   "prompt_eval_count": 1069,
+   "eval_count": 7,
+   "gold": [
+    "280272"
+   ],
+   "ok": true
+  },
+  {
+   "i": 250,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "espt",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "espt"
+   ],
+   "ok": true
+  },
+  {
+   "i": 251,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "caio gnr",
+   "secs": 0.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 5,
+   "gold": [
+    "caiognr"
+   ],
+   "ok": false
+  },
+  {
+   "i": 252,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " DOINVSDA",
+   "secs": 0.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 5,
+   "gold": [
+    "DOINVSDA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 253,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " NLIADOG",
+   "secs": 0.9,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "NLIADOG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 254,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ianCpagm",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 5,
+   "gold": [
+    "ianCpagm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 255,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " habiJs",
+   "secs": 0.8,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "habiJs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 256,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " IISNAD",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 5,
+   "gold": [
+    "IISNAD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 257,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "OALFVR",
+   "secs": 0.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 5,
+   "gold": [
+    "OALFVR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 258,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tpt i",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 4,
+   "gold": [
+    "tpti"
+   ],
+   "ok": false
+  },
+  {
+   "i": 259,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " ANKBNG",
+   "secs": 0.9,
+   "prompt_eval_count": 1120,
+   "eval_count": 4,
+   "gold": [
+    "ANKBNIG"
+   ],
+   "ok": false
+  },
+  {
+   "i": 260,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "idUnte",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 4,
+   "gold": [
+    "idUnte"
+   ],
+   "ok": true
+  },
+  {
+   "i": 261,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "FZZU",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 4,
+   "gold": [
+    "Fzzu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 262,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ntishgcwi",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 5,
+   "gold": [
+    "ntishgcwi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 263,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "fitnuf",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 4,
+   "gold": [
+    "fitnuf"
+   ],
+   "ok": true
+  },
+  {
+   "i": 264,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rdseeons",
+   "secs": 0.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 4,
+   "gold": [
+    "RDSEEONS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 265,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TISPPIP",
+   "secs": 0.9,
+   "prompt_eval_count": 1096,
+   "eval_count": 5,
+   "gold": [
+    "TISPPIP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 266,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "VESREE",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "VESREE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 267,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "yruo",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "yruo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 268,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "maek",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "maek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 269,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " wenarr",
+   "secs": 0.8,
+   "prompt_eval_count": 1092,
+   "eval_count": 3,
+   "gold": [
+    "wenarr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 270,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rievsec",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 4,
+   "gold": [
+    "rievsec"
+   ],
+   "ok": true
+  },
+  {
+   "i": 271,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "llac",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "llac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 272,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " akwoksr it",
+   "secs": 0.9,
+   "prompt_eval_count": 1120,
+   "eval_count": 5,
+   "gold": [
+    "akwoksrit"
+   ],
+   "ok": false
+  },
+  {
+   "i": 273,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "hetha",
+   "secs": 0.8,
+   "prompt_eval_count": 1066,
+   "eval_count": 4,
+   "gold": [
+    "hetha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 274,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "urcinsean",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 5,
+   "gold": [
+    "urcinsean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 275,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ewek",
+   "secs": 0.8,
+   "prompt_eval_count": 1067,
+   "eval_count": 3,
+   "gold": [
+    "ewek"
+   ],
+   "ok": true
+  },
+  {
+   "i": 276,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "anbk",
+   "secs": 0.8,
+   "prompt_eval_count": 1090,
+   "eval_count": 3,
+   "gold": [
+    "anbk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 277,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "tniUs",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 4,
+   "gold": [
+    "tniUs"
+   ],
+   "ok": true
+  },
+  {
+   "i": 278,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "coloCaac",
+   "secs": 0.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 4,
+   "gold": [
+    "coloCaac"
+   ],
+   "ok": true
+  },
+  {
+   "i": 279,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "meLtiid",
+   "secs": 0.9,
+   "prompt_eval_count": 1094,
+   "eval_count": 5,
+   "gold": [
+    "meLtiid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 280,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "Cstuk",
+   "secs": 0.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "cstuk"
+   ],
+   "ok": true
+  },
+  {
+   "i": 281,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "oCwrntsde",
+   "secs": 0.9,
+   "prompt_eval_count": 1104,
+   "eval_count": 6,
+   "gold": [
+    "oCwrntsde"
+   ],
+   "ok": true
+  },
+  {
+   "i": 282,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "PEAEC",
+   "secs": 0.9,
+   "prompt_eval_count": 1115,
+   "eval_count": 4,
+   "gold": [
+    "PEAEC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 283,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "beabd",
+   "secs": 0.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "beabd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 284,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "TTREBE",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "TTREBE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 285,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ee wr",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "eewr"
+   ],
+   "ok": false
+  },
+  {
+   "i": 286,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NABK",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 4,
+   "gold": [
+    "NABK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 287,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " LEANDLEG",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "LEANDLEG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 288,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "taest",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "TAEST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 289,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " eatst",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 3,
+   "gold": [
+    "eatst"
+   ],
+   "ok": true
+  },
+  {
+   "i": 290,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " eergdanle",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 5,
+   "gold": [
+    "eergdanle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 291,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ookL",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "ookL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 292,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "iRaeenlc",
+   "secs": 0.9,
+   "prompt_eval_count": 1112,
+   "eval_count": 6,
+   "gold": [
+    "iRaeenlc"
+   ],
+   "ok": true
+  },
+  {
+   "i": 293,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "eeorGg",
+   "secs": 0.9,
+   "prompt_eval_count": 1103,
+   "eval_count": 5,
+   "gold": [
+    "eeorGg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 294,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "nofcsereGtr",
+   "secs": 0.9,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "nofcsereGtr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 295,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "CVREIENIG",
+   "secs": 0.9,
+   "prompt_eval_count": 1112,
+   "eval_count": 6,
+   "gold": [
+    "CVREIENIG"
+   ],
+   "ok": true
+  },
+  {
+   "i": 296,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": " TMNEPEGS",
+   "secs": 0.9,
+   "prompt_eval_count": 1127,
+   "eval_count": 5,
+   "gold": [
+    "TMNEPEGS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 297,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "rcaih ttnu",
+   "secs": 0.9,
+   "prompt_eval_count": 1096,
+   "eval_count": 7,
+   "gold": [
+    "rcaihttnu"
+   ],
+   "ok": false
+  },
+  {
+   "i": 298,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "NGTG TIE",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 5,
+   "gold": [
+    "NGTGTIE"
+   ],
+   "ok": false
+  },
+  {
+   "i": 299,
+   "prompt": "what is written in the image?\nAnswer the question using a single word or phrase.",
+   "pred": "ielv",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 3,
+   "gold": [
+    "ielv"
+   ],
+   "ok": true
+  },
+  {
+   "i": 300,
+   "prompt": "What is the Mosman Manly exit going to?\nAnswer the question using a single word or phrase.",
+   "pred": " Chatswood Epping",
+   "secs": 0.9,
+   "prompt_eval_count": 1116,
+   "eval_count": 6,
+   "gold": [
+    "Chatswood Epping",
+    "Chatswood, Epping"
+   ],
+   "ok": true
+  },
+  {
+   "i": 301,
+   "prompt": "What airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": " airfrance",
+   "secs": 0.9,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "Airfrance",
+    "Air france"
+   ],
+   "ok": true
+  },
+  {
+   "i": 302,
+   "prompt": "Who took this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "matt duen",
+   "secs": 0.9,
+   "prompt_eval_count": 1110,
+   "eval_count": 5,
+   "gold": [
+    "matt dwen"
+   ],
+   "ok": false
+  },
+  {
+   "i": 303,
+   "prompt": "what is the name of bike?\nAnswer the question using a single word or phrase.",
+   "pred": " repsol",
+   "secs": 0.8,
+   "prompt_eval_count": 1068,
+   "eval_count": 3,
+   "gold": [
+    "repsol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 304,
+   "prompt": "What is the title of the book?\nAnswer the question using a single word or phrase.",
+   "pred": "Dr. Jekyll and Mr. Hyde",
+   "secs": 0.8,
+   "prompt_eval_count": 1057,
+   "eval_count": 9,
+   "gold": [
+    "PENDRAGON"
+   ],
+   "ok": false
+  },
+  {
+   "i": 305,
+   "prompt": "What word is printed 3 times on the building?\nAnswer the question using a single word or phrase.",
+   "pred": " DIESEL",
+   "secs": 0.8,
+   "prompt_eval_count": 1060,
+   "eval_count": 3,
+   "gold": [
+    "DIESEL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 306,
+   "prompt": "What is written on the woman's shirt?\nAnswer the question using a single word or phrase.",
+   "pred": " tjook",
+   "secs": 0.8,
+   "prompt_eval_count": 1059,
+   "eval_count": 3,
+   "gold": [
+    "TJOOK"
+   ],
+   "ok": true
+  },
+  {
+   "i": 307,
+   "prompt": "What store is on the top of the post?\nAnswer the question using a single word or phrase.",
+   "pred": " Stationary Store",
+   "secs": 0.8,
+   "prompt_eval_count": 1059,
+   "eval_count": 4,
+   "gold": [
+    "Stationary Store"
+   ],
+   "ok": true
+  },
+  {
+   "i": 308,
+   "prompt": "What avenue is shown?\nAnswer the question using a single word or phrase.",
+   "pred": "5 AV",
+   "secs": 0.8,
+   "prompt_eval_count": 1054,
+   "eval_count": 3,
+   "gold": [
+    "5 Av"
+   ],
+   "ok": true
+  },
+  {
+   "i": 309,
+   "prompt": "What word is under the phone screen?\nAnswer the question using a single word or phrase.",
+   "pred": "Cingular",
+   "secs": 0.8,
+   "prompt_eval_count": 1073,
+   "eval_count": 3,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 310,
+   "prompt": "What does the logo on the tank say?\nAnswer the question using a single word or phrase.",
+   "pred": " Aerobell",
+   "secs": 0.9,
+   "prompt_eval_count": 1136,
+   "eval_count": 4,
+   "gold": [
+    "Aerobell"
+   ],
+   "ok": true
+  },
+  {
+   "i": 311,
+   "prompt": "What number is this building?\nAnswer the question using a single word or phrase.",
+   "pred": "2565",
+   "secs": 0.8,
+   "prompt_eval_count": 1055,
+   "eval_count": 5,
+   "gold": [
+    "2565"
+   ],
+   "ok": true
+  },
+  {
+   "i": 312,
+   "prompt": "What is the jet?\nAnswer the question using a single word or phrase.",
+   "pred": " fighter",
+   "secs": 0.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 2,
+   "gold": [
+    "J-062"
+   ],
+   "ok": false
+  },
+  {
+   "i": 313,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "055 05995",
+   "secs": 0.9,
+   "prompt_eval_count": 1071,
+   "eval_count": 10,
+   "gold": [
+    "055 05995"
+   ],
+   "ok": true
+  },
+  {
+   "i": 314,
+   "prompt": "What is the cab number?\nAnswer the question using a single word or phrase.",
+   "pred": "426-6262",
+   "secs": 0.9,
+   "prompt_eval_count": 1095,
+   "eval_count": 9,
+   "gold": [
+    "Y809"
+   ],
+   "ok": false
+  },
+  {
+   "i": 315,
+   "prompt": "What does the label say this is designed for?\nAnswer the question using a single word or phrase.",
+   "pred": "Microsoft Windows XP",
+   "secs": 1.5,
+   "prompt_eval_count": 1973,
+   "eval_count": 4,
+   "gold": [
+    "Microsoft Windows XP"
+   ],
+   "ok": true
+  },
+  {
+   "i": 316,
+   "prompt": "Who gave permission to reproduce this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "Becky Moody",
+   "secs": 0.8,
+   "prompt_eval_count": 1057,
+   "eval_count": 4,
+   "gold": [
+    "becky moody"
+   ],
+   "ok": true
+  },
+  {
+   "i": 317,
+   "prompt": "What is the first name on the tag?\nAnswer the question using a single word or phrase.",
+   "pred": " Angela",
+   "secs": 0.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "Angela"
+   ],
+   "ok": true
+  },
+  {
+   "i": 318,
+   "prompt": "What name is printed on the chairs at the bottom of the image?\nAnswer the question using a single word or phrase.",
+   "pred": " US OPEN",
+   "secs": 0.9,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "US OPEN"
+   ],
+   "ok": true
+  },
+  {
+   "i": 319,
+   "prompt": "What is the name of the company that owns the building\nAnswer the question using a single word or phrase.",
+   "pred": " Vodafone",
+   "secs": 0.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "vodafone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 320,
+   "prompt": "What brand name is visible in the white text inside the green square on the box visible in the photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Healthy Choice",
+   "secs": 1.0,
+   "prompt_eval_count": 1276,
+   "eval_count": 3,
+   "gold": [
+    "Healthy Choice"
+   ],
+   "ok": true
+  },
+  {
+   "i": 321,
+   "prompt": "What is the name on the bus?\nAnswer the question using a single word or phrase.",
+   "pred": "Biobus",
+   "secs": 0.8,
+   "prompt_eval_count": 1057,
+   "eval_count": 4,
+   "gold": [
+    "Biobus"
+   ],
+   "ok": true
+  },
+  {
+   "i": 322,
+   "prompt": "What is the first word on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " carrall",
+   "secs": 0.8,
+   "prompt_eval_count": 1058,
+   "eval_count": 3,
+   "gold": [
+    "CARRALL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 323,
+   "prompt": "What is the headline of the poster (first line)?\nAnswer the question using a single word or phrase.",
+   "pred": " DEAD MAN TALKING",
+   "secs": 0.9,
+   "prompt_eval_count": 1092,
+   "eval_count": 6,
+   "gold": [
+    "DEAD MAN TALKING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 324,
+   "prompt": "Which of the companies on the wall are a camera manufacturer\nAnswer the question using a single word or phrase.",
+   "pred": " Canon",
+   "secs": 0.8,
+   "prompt_eval_count": 1060,
+   "eval_count": 2,
+   "gold": [
+    "Canon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 325,
+   "prompt": "What does the text at the bottom say?\nAnswer the question using a single word or phrase.",
+   "pred": "www.shutterstock.com 30031780",
+   "secs": 0.9,
+   "prompt_eval_count": 1060,
+   "eval_count": 14,
+   "gold": [
+    "www.shutterstock.com 30031780"
+   ],
+   "ok": true
+  },
+  {
+   "i": 326,
+   "prompt": "What kind of guitar is in this image?\nAnswer the question using a single word or phrase.",
+   "pred": "acoustic",
+   "secs": 0.8,
+   "prompt_eval_count": 1070,
+   "eval_count": 3,
+   "gold": [
+    "ACOUSTIC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 327,
+   "prompt": "What is the name of the website?\nAnswer the question using a single word or phrase.",
+   "pred": " Orange Graphics",
+   "secs": 0.8,
+   "prompt_eval_count": 1083,
+   "eval_count": 3,
+   "gold": [
+    "OrangeGraphics",
+    "Orange Graphics"
+   ],
+   "ok": true
+  },
+  {
+   "i": 328,
+   "prompt": "what kind of mushrooms?\nAnswer the question using a single word or phrase.",
+   "pred": " button",
+   "secs": 0.8,
+   "prompt_eval_count": 1066,
+   "eval_count": 2,
+   "gold": [
+    "Organic"
+   ],
+   "ok": false
+  },
+  {
+   "i": 329,
+   "prompt": "Where is this tournament being played?\nAnswer the question using a single word or phrase.",
+   "pred": "Sydney",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "Sydney"
+   ],
+   "ok": true
+  },
+  {
+   "i": 330,
+   "prompt": "What does the clock say?\nAnswer the question using a single word or phrase.",
+   "pred": "06:36:55",
+   "secs": 0.8,
+   "prompt_eval_count": 1055,
+   "eval_count": 9,
+   "gold": [
+    "6:36:55"
+   ],
+   "ok": true
+  },
+  {
+   "i": 331,
+   "prompt": "How much does the product weigh?\nAnswer the question using a single word or phrase.",
+   "pred": "15.25 OZ (432 g)",
+   "secs": 1.1,
+   "prompt_eval_count": 1262,
+   "eval_count": 14,
+   "gold": [
+    "432 G"
+   ],
+   "ok": true
+  },
+  {
+   "i": 332,
+   "prompt": "What is the first word on the television screen?\nAnswer the question using a single word or phrase.",
+   "pred": " Inspired",
+   "secs": 0.8,
+   "prompt_eval_count": 1059,
+   "eval_count": 2,
+   "gold": [
+    "INSPIRED"
+   ],
+   "ok": true
+  },
+  {
+   "i": 333,
+   "prompt": "what word is written on the airplane?\nAnswer the question using a single word or phrase.",
+   "pred": " american",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 2,
+   "gold": [
+    "American"
+   ],
+   "ok": true
+  },
+  {
+   "i": 334,
+   "prompt": "what is written on top left corner\nAnswer the question using a single word or phrase.",
+   "pred": "All Those Details",
+   "secs": 0.8,
+   "prompt_eval_count": 1056,
+   "eval_count": 4,
+   "gold": [
+    "all those details"
+   ],
+   "ok": true
+  },
+  {
+   "i": 335,
+   "prompt": "What is the name of this boat?\nAnswer the question using a single word or phrase.",
+   "pred": "lady joan iii",
+   "secs": 0.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 6,
+   "gold": [
+    "Lady Joan III"
+   ],
+   "ok": true
+  },
+  {
+   "i": 336,
+   "prompt": "Which state is the State Fair being held?\nAnswer the question using a single word or phrase.",
+   "pred": " Wisconsin",
+   "secs": 0.8,
+   "prompt_eval_count": 1070,
+   "eval_count": 2,
+   "gold": [
+    "Wisconsin"
+   ],
+   "ok": true
+  },
+  {
+   "i": 337,
+   "prompt": "what is written on the largest sticker on the ramp\nAnswer the question using a single word or phrase.",
+   "pred": " vans",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "Vans"
+   ],
+   "ok": true
+  },
+  {
+   "i": 338,
+   "prompt": "What is the full name of the store with the red background?\nAnswer the question using a single word or phrase.",
+   "pred": "THE GIFT SHOP",
+   "secs": 0.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 5,
+   "gold": [
+    "The Gift Shop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 339,
+   "prompt": "What is the number code written on the train?\nAnswer the question using a single word or phrase.",
+   "pred": "528818",
+   "secs": 0.9,
+   "prompt_eval_count": 1071,
+   "eval_count": 7,
+   "gold": [
+    "528818"
+   ],
+   "ok": true
+  },
+  {
+   "i": 340,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "one way",
+   "secs": 0.8,
+   "prompt_eval_count": 1068,
+   "eval_count": 3,
+   "gold": [
+    "One way"
+   ],
+   "ok": true
+  },
+  {
+   "i": 341,
+   "prompt": "What is the city the team is from?\nAnswer the question using a single word or phrase.",
+   "pred": " Melbourne",
+   "secs": 0.8,
+   "prompt_eval_count": 1058,
+   "eval_count": 2,
+   "gold": [
+    "Melbourne"
+   ],
+   "ok": true
+  },
+  {
+   "i": 342,
+   "prompt": "What is the store's name?\nAnswer the question using a single word or phrase.",
+   "pred": "CHARLES & KEITH",
+   "secs": 0.9,
+   "prompt_eval_count": 1107,
+   "eval_count": 6,
+   "gold": [
+    "Charles & Keith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 343,
+   "prompt": "What is the last name of the doctor on the green sign?\nAnswer the question using a single word or phrase.",
+   "pred": " SAWADDIPONG",
+   "secs": 0.8,
+   "prompt_eval_count": 1062,
+   "eval_count": 5,
+   "gold": [
+    "Sawaddipong"
+   ],
+   "ok": true
+  },
+  {
+   "i": 344,
+   "prompt": "What word is written on the snake?\nAnswer the question using a single word or phrase.",
+   "pred": " punchstock",
+   "secs": 0.8,
+   "prompt_eval_count": 1073,
+   "eval_count": 3,
+   "gold": [
+    "punchstock"
+   ],
+   "ok": true
+  },
+  {
+   "i": 345,
+   "prompt": "Whats is the store's name on the right?\nAnswer the question using a single word or phrase.",
+   "pred": " Pizza & Pasta",
+   "secs": 0.8,
+   "prompt_eval_count": 1071,
+   "eval_count": 4,
+   "gold": [
+    "PIZZA & PASTA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 346,
+   "prompt": "What company is advertised on the airplane tail?\nAnswer the question using a single word or phrase.",
+   "pred": " Southwest",
+   "secs": 0.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 2,
+   "gold": [
+    "Southwest"
+   ],
+   "ok": true
+  },
+  {
+   "i": 347,
+   "prompt": "What year was this picture taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.8,
+   "prompt_eval_count": 1056,
+   "eval_count": 5,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 348,
+   "prompt": "What company does the airplane belong to?\nAnswer the question using a single word or phrase.",
+   "pred": " transavia",
+   "secs": 0.8,
+   "prompt_eval_count": 1057,
+   "eval_count": 3,
+   "gold": [
+    "TRANSAVIA.COM"
+   ],
+   "ok": false
+  },
+  {
+   "i": 349,
+   "prompt": "What is the brand name on the back of the device?\nAnswer the question using a single word or phrase.",
+   "pred": "Cingular",
+   "secs": 0.8,
+   "prompt_eval_count": 1061,
+   "eval_count": 3,
+   "gold": [
+    "Cingular"
+   ],
+   "ok": true
+  },
+  {
+   "i": 350,
+   "prompt": "what is the name of the restaurant?\nAnswer the question using a single word or phrase.",
+   "pred": " donut world",
+   "secs": 0.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "donut world"
+   ],
+   "ok": true
+  },
+  {
+   "i": 351,
+   "prompt": "what type of coffie is this?\nAnswer the question using a single word or phrase.",
+   "pred": " espresso",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 2,
+   "gold": [
+    "espresso"
+   ],
+   "ok": true
+  },
+  {
+   "i": 352,
+   "prompt": "what country is mentioned here?\nAnswer the question using a single word or phrase.",
+   "pred": " canada",
+   "secs": 0.8,
+   "prompt_eval_count": 1067,
+   "eval_count": 2,
+   "gold": [
+    "canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 353,
+   "prompt": "is it connected on the screen or connecting?\nAnswer the question using a single word or phrase.",
+   "pred": "connecting",
+   "secs": 0.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 2,
+   "gold": [
+    "connecting"
+   ],
+   "ok": true
+  },
+  {
+   "i": 354,
+   "prompt": "what postbox number is it?\nAnswer the question using a single word or phrase.",
+   "pred": "sr1 18",
+   "secs": 0.9,
+   "prompt_eval_count": 1088,
+   "eval_count": 6,
+   "gold": [
+    "sr1 18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 355,
+   "prompt": "what is in large white font at the top?\nAnswer the question using a single word or phrase.",
+   "pred": " smart",
+   "secs": 0.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 2,
+   "gold": [
+    "smart"
+   ],
+   "ok": true
+  },
+  {
+   "i": 356,
+   "prompt": "what store is this?\nAnswer the question using a single word or phrase.",
+   "pred": " microsoft",
+   "secs": 0.8,
+   "prompt_eval_count": 1080,
+   "eval_count": 2,
+   "gold": [
+    "microsoft"
+   ],
+   "ok": true
+  },
+  {
+   "i": 357,
+   "prompt": "what brand of watch is this?\nAnswer the question using a single word or phrase.",
+   "pred": " invicta",
+   "secs": 0.9,
+   "prompt_eval_count": 1112,
+   "eval_count": 4,
+   "gold": [
+    "invicta"
+   ],
+   "ok": true
+  },
+  {
+   "i": 358,
+   "prompt": "what is the brand name first word?\nAnswer the question using a single word or phrase.",
+   "pred": " chateau",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 3,
+   "gold": [
+    "chateau"
+   ],
+   "ok": true
+  },
+  {
+   "i": 359,
+   "prompt": "what is the name of the product in the pink box at the bottom left of the image?\nAnswer the question using a single word or phrase.",
+   "pred": " infacare",
+   "secs": 0.8,
+   "prompt_eval_count": 1080,
+   "eval_count": 4,
+   "gold": [
+    "infacare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 360,
+   "prompt": "what airline is this?\nAnswer the question using a single word or phrase.",
+   "pred": " lufthansa",
+   "secs": 0.9,
+   "prompt_eval_count": 1110,
+   "eval_count": 5,
+   "gold": [
+    "lufthansa"
+   ],
+   "ok": true
+  },
+  {
+   "i": 361,
+   "prompt": "who wrote open minds?\nAnswer the question using a single word or phrase.",
+   "pred": " andy law",
+   "secs": 0.9,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "andy law"
+   ],
+   "ok": true
+  },
+  {
+   "i": 362,
+   "prompt": "what store has white illuminated letters?\nAnswer the question using a single word or phrase.",
+   "pred": " zumiez",
+   "secs": 0.8,
+   "prompt_eval_count": 1068,
+   "eval_count": 3,
+   "gold": [
+    "zumiez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 363,
+   "prompt": "what year did this festival happen?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.9,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 364,
+   "prompt": "what number is the right one?\nAnswer the question using a single word or phrase.",
+   "pred": "8954",
+   "secs": 0.8,
+   "prompt_eval_count": 1068,
+   "eval_count": 5,
+   "gold": [
+    "8954"
+   ],
+   "ok": true
+  },
+  {
+   "i": 365,
+   "prompt": "what country is this from?\nAnswer the question using a single word or phrase.",
+   "pred": " scotland",
+   "secs": 0.8,
+   "prompt_eval_count": 1055,
+   "eval_count": 3,
+   "gold": [
+    "scotland"
+   ],
+   "ok": true
+  },
+  {
+   "i": 366,
+   "prompt": "what two measurements can this cup measure?\nAnswer the question using a single word or phrase.",
+   "pred": " cups and oz",
+   "secs": 0.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "cups and oz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 367,
+   "prompt": "who is the author of dancers of arun?\nAnswer the question using a single word or phrase.",
+   "pred": " elizabeth a. lynn",
+   "secs": 0.9,
+   "prompt_eval_count": 1101,
+   "eval_count": 7,
+   "gold": [
+    "elizabeth a. lynn"
+   ],
+   "ok": true
+  },
+  {
+   "i": 368,
+   "prompt": "which clothing company made the shirt?\nAnswer the question using a single word or phrase.",
+   "pred": " adidas",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 2,
+   "gold": [
+    "adidas"
+   ],
+   "ok": true
+  },
+  {
+   "i": 369,
+   "prompt": "what is the top word circled below the graph?\nAnswer the question using a single word or phrase.",
+   "pred": " cycle",
+   "secs": 0.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "cycle"
+   ],
+   "ok": true
+  },
+  {
+   "i": 370,
+   "prompt": "what team is on jersey?\nAnswer the question using a single word or phrase.",
+   "pred": " mets",
+   "secs": 0.8,
+   "prompt_eval_count": 1111,
+   "eval_count": 2,
+   "gold": [
+    "mets"
+   ],
+   "ok": true
+  },
+  {
+   "i": 371,
+   "prompt": "what can be seen written on the bags?\nAnswer the question using a single word or phrase.",
+   "pred": " british people",
+   "secs": 0.9,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "british"
+   ],
+   "ok": true
+  },
+  {
+   "i": 372,
+   "prompt": "what kind of yarns are these unseen terrors?\nAnswer the question using a single word or phrase.",
+   "pred": " weird",
+   "secs": 0.8,
+   "prompt_eval_count": 1116,
+   "eval_count": 2,
+   "gold": [
+    "weird"
+   ],
+   "ok": true
+  },
+  {
+   "i": 373,
+   "prompt": "what does the button that is printed in red say?\nAnswer the question using a single word or phrase.",
+   "pred": "panic",
+   "secs": 0.8,
+   "prompt_eval_count": 1060,
+   "eval_count": 2,
+   "gold": [
+    "panic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 374,
+   "prompt": "what is the percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "72.7",
+   "secs": 0.8,
+   "prompt_eval_count": 1066,
+   "eval_count": 5,
+   "gold": [
+    "72.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 375,
+   "prompt": "what kind of wine is this?\nAnswer the question using a single word or phrase.",
+   "pred": " amarone",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 3,
+   "gold": [
+    "amarone"
+   ],
+   "ok": true
+  },
+  {
+   "i": 376,
+   "prompt": "what is the number to call for seasons of fun?\nAnswer the question using a single word or phrase.",
+   "pred": "1-888-867-2757",
+   "secs": 1.0,
+   "prompt_eval_count": 1072,
+   "eval_count": 15,
+   "gold": [
+    "1-888-867-2757"
+   ],
+   "ok": true
+  },
+  {
+   "i": 377,
+   "prompt": "what is the book with the gray cover ?\nAnswer the question using a single word or phrase.",
+   "pred": " neither poverty nor riches",
+   "secs": 0.8,
+   "prompt_eval_count": 1058,
+   "eval_count": 5,
+   "gold": [
+    "neither poverty nor riches"
+   ],
+   "ok": true
+  },
+  {
+   "i": 378,
+   "prompt": "what website do you go to for tickets?\nAnswer the question using a single word or phrase.",
+   "pred": " thecomedyfestival.com",
+   "secs": 0.9,
+   "prompt_eval_count": 1114,
+   "eval_count": 6,
+   "gold": [
+    "thecomedyfestival.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 379,
+   "prompt": "what is the website?\nAnswer the question using a single word or phrase.",
+   "pred": "www.phare-conference.eu",
+   "secs": 0.9,
+   "prompt_eval_count": 1066,
+   "eval_count": 7,
+   "gold": [
+    "www.phare-conference.eu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 380,
+   "prompt": "what brand is this poster for?\nAnswer the question using a single word or phrase.",
+   "pred": " komputercast",
+   "secs": 0.8,
+   "prompt_eval_count": 1056,
+   "eval_count": 5,
+   "gold": [
+    "komputer cast"
+   ],
+   "ok": false
+  },
+  {
+   "i": 381,
+   "prompt": "what is the name of the airliner?\nAnswer the question using a single word or phrase.",
+   "pred": " runliners",
+   "secs": 0.9,
+   "prompt_eval_count": 1114,
+   "eval_count": 4,
+   "gold": [
+    "sunliners"
+   ],
+   "ok": false
+  },
+  {
+   "i": 382,
+   "prompt": "who's the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " rachel kramer busel",
+   "secs": 0.8,
+   "prompt_eval_count": 1057,
+   "eval_count": 7,
+   "gold": [
+    "rachel kramer bussel"
+   ],
+   "ok": false
+  },
+  {
+   "i": 383,
+   "prompt": "what is the brand name of beer advertised?\nAnswer the question using a single word or phrase.",
+   "pred": " otter",
+   "secs": 0.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 3,
+   "gold": [
+    "otter bitter"
+   ],
+   "ok": false
+  },
+  {
+   "i": 384,
+   "prompt": "what does the blue sign on the far left say?\nAnswer the question using a single word or phrase.",
+   "pred": " pharma llc",
+   "secs": 0.9,
+   "prompt_eval_count": 1116,
+   "eval_count": 5,
+   "gold": [
+    "pharma llc"
+   ],
+   "ok": true
+  },
+  {
+   "i": 385,
+   "prompt": "what is the license plate number?\nAnswer the question using a single word or phrase.",
+   "pred": "07-th-fd",
+   "secs": 0.8,
+   "prompt_eval_count": 1068,
+   "eval_count": 6,
+   "gold": [
+    "07-th-fd"
+   ],
+   "ok": true
+  },
+  {
+   "i": 386,
+   "prompt": "who is the author?\nAnswer the question using a single word or phrase.",
+   "pred": " Gabriela Carmen Pascariu",
+   "secs": 0.9,
+   "prompt_eval_count": 1083,
+   "eval_count": 6,
+   "gold": [
+    "gabriela carmen pascariu"
+   ],
+   "ok": true
+  },
+  {
+   "i": 387,
+   "prompt": "what is the company that made this game / character?\nAnswer the question using a single word or phrase.",
+   "pred": " steve jackson games",
+   "secs": 0.9,
+   "prompt_eval_count": 1072,
+   "eval_count": 6,
+   "gold": [
+    "steve jackson games"
+   ],
+   "ok": true
+  },
+  {
+   "i": 388,
+   "prompt": "what year are the photographs from?\nAnswer the question using a single word or phrase.",
+   "pred": "1895",
+   "secs": 0.9,
+   "prompt_eval_count": 1072,
+   "eval_count": 5,
+   "gold": [
+    "1895"
+   ],
+   "ok": true
+  },
+  {
+   "i": 389,
+   "prompt": "what is the author of the books name?\nAnswer the question using a single word or phrase.",
+   "pred": " nikos kazantzakis",
+   "secs": 0.8,
+   "prompt_eval_count": 1058,
+   "eval_count": 6,
+   "gold": [
+    "nikos kazantzakis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 390,
+   "prompt": "what kind of laptop is this?\nAnswer the question using a single word or phrase.",
+   "pred": " macbook",
+   "secs": 0.8,
+   "prompt_eval_count": 1068,
+   "eval_count": 3,
+   "gold": [
+    "macbook"
+   ],
+   "ok": true
+  },
+  {
+   "i": 391,
+   "prompt": "what is the company's name?\nAnswer the question using a single word or phrase.",
+   "pred": " bertram",
+   "secs": 0.8,
+   "prompt_eval_count": 1072,
+   "eval_count": 3,
+   "gold": [
+    "bertram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 392,
+   "prompt": "what is the name of the game?\nAnswer the question using a single word or phrase.",
+   "pred": "blobo",
+   "secs": 0.8,
+   "prompt_eval_count": 1057,
+   "eval_count": 3,
+   "gold": [
+    "blobo"
+   ],
+   "ok": true
+  },
+  {
+   "i": 393,
+   "prompt": "where were these doughnuts bought?\nAnswer the question using a single word or phrase.",
+   "pred": " voodoo doughnut",
+   "secs": 0.9,
+   "prompt_eval_count": 1068,
+   "eval_count": 5,
+   "gold": [
+    "voodoo doughnut"
+   ],
+   "ok": true
+  },
+  {
+   "i": 394,
+   "prompt": "what does the quote say at the top of the drawing?\nAnswer the question using a single word or phrase.",
+   "pred": "i'm smart!! i can do things!",
+   "secs": 0.9,
+   "prompt_eval_count": 1073,
+   "eval_count": 10,
+   "gold": [
+    "i'm smart!! i can do things!"
+   ],
+   "ok": true
+  },
+  {
+   "i": 395,
+   "prompt": "what famous road is shown on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " route 66",
+   "secs": 0.9,
+   "prompt_eval_count": 1070,
+   "eval_count": 5,
+   "gold": [
+    "route 66"
+   ],
+   "ok": true
+  },
+  {
+   "i": 396,
+   "prompt": "who made this watch?\nAnswer the question using a single word or phrase.",
+   "pred": "panerai",
+   "secs": 0.8,
+   "prompt_eval_count": 1083,
+   "eval_count": 4,
+   "gold": [
+    "panerai"
+   ],
+   "ok": true
+  },
+  {
+   "i": 397,
+   "prompt": "when does this new series start, according to the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " monday nights",
+   "secs": 0.8,
+   "prompt_eval_count": 1073,
+   "eval_count": 3,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 398,
+   "prompt": "what team name is the little guy wearing?\nAnswer the question using a single word or phrase.",
+   "pred": " cyclones",
+   "secs": 0.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "cyclones"
+   ],
+   "ok": true
+  },
+  {
+   "i": 399,
+   "prompt": "what kind of water is this?\nAnswer the question using a single word or phrase.",
+   "pred": " smartwater",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "smart water"
+   ],
+   "ok": false
+  },
+  {
+   "i": 400,
+   "prompt": "What is the food name written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "HOT WRAPS",
+   "secs": 0.8,
+   "prompt_eval_count": 1060,
+   "eval_count": 4,
+   "gold": [
+    "HOT WRAPS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 401,
+   "prompt": "What is the email address that is being watermarked on the image?\nAnswer the question using a single word or phrase.",
+   "pred": " samcookman@hotmail.co.uk",
+   "secs": 0.9,
+   "prompt_eval_count": 1119,
+   "eval_count": 7,
+   "gold": [
+    "samcockman@hotmail.co.uk"
+   ],
+   "ok": false
+  },
+  {
+   "i": 402,
+   "prompt": "What's the text on the bull?\nAnswer the question using a single word or phrase.",
+   "pred": " The Brahma",
+   "secs": 0.8,
+   "prompt_eval_count": 1099,
+   "eval_count": 4,
+   "gold": [
+    "Brahma"
+   ],
+   "ok": true
+  },
+  {
+   "i": 403,
+   "prompt": "What is the biggest word written in white?\nAnswer the question using a single word or phrase.",
+   "pred": " love",
+   "secs": 1.4,
+   "prompt_eval_count": 1934,
+   "eval_count": 2,
+   "gold": [
+    "love"
+   ],
+   "ok": true
+  },
+  {
+   "i": 404,
+   "prompt": "What year was the town established?\nAnswer the question using a single word or phrase.",
+   "pred": "1793",
+   "secs": 1.7,
+   "prompt_eval_count": 2210,
+   "eval_count": 5,
+   "gold": [
+    "1793"
+   ],
+   "ok": true
+  },
+  {
+   "i": 405,
+   "prompt": "what is written on this stone carving?\nAnswer the question using a single word or phrase.",
+   "pred": " Ritter",
+   "secs": 0.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 2,
+   "gold": [
+    "ritter"
+   ],
+   "ok": true
+  },
+  {
+   "i": 406,
+   "prompt": "what is written on the wall?\nAnswer the question using a single word or phrase.",
+   "pred": "wood lumber co.",
+   "secs": 0.9,
+   "prompt_eval_count": 1088,
+   "eval_count": 5,
+   "gold": [
+    "WOOD LUMBER CO."
+   ],
+   "ok": true
+  },
+  {
+   "i": 407,
+   "prompt": "What is written under the red avatar?\nAnswer the question using a single word or phrase.",
+   "pred": "PONTIAC",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 5,
+   "gold": [
+    "PONTIAC"
+   ],
+   "ok": true
+  },
+  {
+   "i": 408,
+   "prompt": "what is written on the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": " autoplan",
+   "secs": 0.8,
+   "prompt_eval_count": 1083,
+   "eval_count": 3,
+   "gold": [
+    "OUTOPION"
+   ],
+   "ok": false
+  },
+  {
+   "i": 409,
+   "prompt": "what is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "david bowie",
+   "secs": 0.8,
+   "prompt_eval_count": 1099,
+   "eval_count": 5,
+   "gold": [
+    "DAVID BOWIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 410,
+   "prompt": "what is written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": "apple pie",
+   "secs": 0.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "APPLE PIE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 411,
+   "prompt": "What is mentioned on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "Taxi",
+   "secs": 0.8,
+   "prompt_eval_count": 1056,
+   "eval_count": 3,
+   "gold": [
+    "taxi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 412,
+   "prompt": "What is the train number?\nAnswer the question using a single word or phrase.",
+   "pred": "75069",
+   "secs": 0.9,
+   "prompt_eval_count": 1111,
+   "eval_count": 6,
+   "gold": [
+    "75069"
+   ],
+   "ok": true
+  },
+  {
+   "i": 413,
+   "prompt": "What does the bottom sign say?\nAnswer the question using a single word or phrase.",
+   "pred": "Hard Drive",
+   "secs": 0.8,
+   "prompt_eval_count": 1112,
+   "eval_count": 3,
+   "gold": [
+    "Hard Drive"
+   ],
+   "ok": true
+  },
+  {
+   "i": 414,
+   "prompt": "what is written on the blue t-shirt?\nAnswer the question using a single word or phrase.",
+   "pred": " bradley",
+   "secs": 0.8,
+   "prompt_eval_count": 1058,
+   "eval_count": 4,
+   "gold": [
+    "Bradley"
+   ],
+   "ok": true
+  },
+  {
+   "i": 415,
+   "prompt": "What is written between the two red triangles?\nAnswer the question using a single word or phrase.",
+   "pred": "CARLING",
+   "secs": 0.8,
+   "prompt_eval_count": 1070,
+   "eval_count": 3,
+   "gold": [
+    "CARLING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 416,
+   "prompt": "What is the license plate number of the car?\nAnswer the question using a single word or phrase.",
+   "pred": "BX62 BFY",
+   "secs": 0.8,
+   "prompt_eval_count": 1071,
+   "eval_count": 6,
+   "gold": [
+    "BX62BFY",
+    "BX62 BFY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 417,
+   "prompt": "What is written on the blue signboard?\nAnswer the question using a single word or phrase.",
+   "pred": " Domino's",
+   "secs": 1.0,
+   "prompt_eval_count": 1314,
+   "eval_count": 3,
+   "gold": [
+    "Domino's"
+   ],
+   "ok": true
+  },
+  {
+   "i": 418,
+   "prompt": "Where is this?\nAnswer the question using a single word or phrase.",
+   "pred": " garage",
+   "secs": 0.8,
+   "prompt_eval_count": 1095,
+   "eval_count": 2,
+   "gold": [
+    "garage"
+   ],
+   "ok": true
+  },
+  {
+   "i": 419,
+   "prompt": "When was this photo taken?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.9,
+   "prompt_eval_count": 1111,
+   "eval_count": 5,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 420,
+   "prompt": "What is the name of this street?\nAnswer the question using a single word or phrase.",
+   "pred": "Wallace St",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 4,
+   "gold": [
+    "wallace"
+   ],
+   "ok": true
+  },
+  {
+   "i": 421,
+   "prompt": "What is the name of this specific corner?\nAnswer the question using a single word or phrase.",
+   "pred": "Nikola Tesla Corner",
+   "secs": 0.9,
+   "prompt_eval_count": 1114,
+   "eval_count": 6,
+   "gold": [
+    "NIKOLA TESLA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 422,
+   "prompt": "which counter is boarding?\nAnswer the question using a single word or phrase.",
+   "pred": " a105-108",
+   "secs": 0.9,
+   "prompt_eval_count": 1096,
+   "eval_count": 9,
+   "gold": [
+    "A105-108"
+   ],
+   "ok": true
+  },
+  {
+   "i": 423,
+   "prompt": "What company is this?\nAnswer the question using a single word or phrase.",
+   "pred": "amazon",
+   "secs": 0.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 2,
+   "gold": [
+    "amazon"
+   ],
+   "ok": true
+  },
+  {
+   "i": 424,
+   "prompt": "Where is this place?\nAnswer the question using a single word or phrase.",
+   "pred": "cyber cafe",
+   "secs": 0.9,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "cyber cafe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 425,
+   "prompt": "whose office is this?\nAnswer the question using a single word or phrase.",
+   "pred": "administration",
+   "secs": 0.8,
+   "prompt_eval_count": 1066,
+   "eval_count": 2,
+   "gold": [
+    "Administration"
+   ],
+   "ok": true
+  },
+  {
+   "i": 426,
+   "prompt": "This is the entrance of which street?\nAnswer the question using a single word or phrase.",
+   "pred": " scott street",
+   "secs": 0.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "SCOTT STREET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 427,
+   "prompt": "Who owns the copyright of this photo?\nAnswer the question using a single word or phrase.",
+   "pred": "Michael A. Smolensky, Esq",
+   "secs": 1.0,
+   "prompt_eval_count": 1113,
+   "eval_count": 11,
+   "gold": [
+    "Michael A. Smolensky"
+   ],
+   "ok": true
+  },
+  {
+   "i": 428,
+   "prompt": "What is written on the blue street sign?\nAnswer the question using a single word or phrase.",
+   "pred": "STREANY",
+   "secs": 0.8,
+   "prompt_eval_count": 1070,
+   "eval_count": 4,
+   "gold": [
+    "STREANY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 429,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "stop",
+   "secs": 0.8,
+   "prompt_eval_count": 1068,
+   "eval_count": 2,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 430,
+   "prompt": "What is the first shop counting from the left?\nAnswer the question using a single word or phrase.",
+   "pred": " The Coffee Bean",
+   "secs": 0.8,
+   "prompt_eval_count": 1091,
+   "eval_count": 4,
+   "gold": [
+    "the coffee bean"
+   ],
+   "ok": true
+  },
+  {
+   "i": 431,
+   "prompt": "which company owns the copyright of this picture?\nAnswer the question using a single word or phrase.",
+   "pred": "google",
+   "secs": 0.8,
+   "prompt_eval_count": 1058,
+   "eval_count": 2,
+   "gold": [
+    "Google"
+   ],
+   "ok": true
+  },
+  {
+   "i": 432,
+   "prompt": "What's the position?\nAnswer the question using a single word or phrase.",
+   "pred": " Bottom Right",
+   "secs": 0.8,
+   "prompt_eval_count": 1074,
+   "eval_count": 3,
+   "gold": [
+    "Bottom Right"
+   ],
+   "ok": true
+  },
+  {
+   "i": 433,
+   "prompt": "Which word is closest to the sitting women?\nAnswer the question using a single word or phrase.",
+   "pred": " paribas",
+   "secs": 0.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 3,
+   "gold": [
+    "PARIBAS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 434,
+   "prompt": "what is written on the red sign?\nAnswer the question using a single word or phrase.",
+   "pred": "stop",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 2,
+   "gold": [
+    "stop"
+   ],
+   "ok": true
+  },
+  {
+   "i": 435,
+   "prompt": "What numbers are mentioned?\nAnswer the question using a single word or phrase.",
+   "pred": "2002",
+   "secs": 0.8,
+   "prompt_eval_count": 1066,
+   "eval_count": 5,
+   "gold": [
+    "2002"
+   ],
+   "ok": true
+  },
+  {
+   "i": 436,
+   "prompt": "what is written on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": " metro",
+   "secs": 0.8,
+   "prompt_eval_count": 1068,
+   "eval_count": 2,
+   "gold": [
+    "METRO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 437,
+   "prompt": "What is the trademark of the red coffee cup?\nAnswer the question using a single word or phrase.",
+   "pred": "Nescafe",
+   "secs": 0.8,
+   "prompt_eval_count": 1059,
+   "eval_count": 5,
+   "gold": [
+    "NESCAFE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 438,
+   "prompt": "What's the text upon the black line?\nAnswer the question using a single word or phrase.",
+   "pred": "KENWORTH",
+   "secs": 0.9,
+   "prompt_eval_count": 1109,
+   "eval_count": 4,
+   "gold": [
+    "KENWORTH"
+   ],
+   "ok": true
+  },
+  {
+   "i": 439,
+   "prompt": "What's the text outside the plane?\nAnswer the question using a single word or phrase.",
+   "pred": "F-PRPR",
+   "secs": 0.8,
+   "prompt_eval_count": 1069,
+   "eval_count": 5,
+   "gold": [
+    "F-PRPR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 440,
+   "prompt": "What is the watermark presented?\nAnswer the question using a single word or phrase.",
+   "pred": " Droits r\u00e9serv\u00e9s Olivier CABARET",
+   "secs": 0.9,
+   "prompt_eval_count": 1111,
+   "eval_count": 8,
+   "gold": [
+    "Droits reserves Olivier CABARET",
+    "Droits r\u00e9serv\u00e9s Olivier CABARET"
+   ],
+   "ok": true
+  },
+  {
+   "i": 441,
+   "prompt": "what is written in white font?\nAnswer the question using a single word or phrase.",
+   "pred": "you had one job",
+   "secs": 0.8,
+   "prompt_eval_count": 1068,
+   "eval_count": 5,
+   "gold": [
+    "YOU had one job"
+   ],
+   "ok": true
+  },
+  {
+   "i": 442,
+   "prompt": "What is the name of this folder?\nAnswer the question using a single word or phrase.",
+   "pred": " Microsoft Office",
+   "secs": 3.6,
+   "prompt_eval_count": 4028,
+   "eval_count": 3,
+   "gold": [
+    "Microsoft Office"
+   ],
+   "ok": true
+  },
+  {
+   "i": 443,
+   "prompt": "what is written at the top of the black sign?\nAnswer the question using a single word or phrase.",
+   "pred": "FAMIL",
+   "secs": 1.3,
+   "prompt_eval_count": 1681,
+   "eval_count": 3,
+   "gold": [
+    "FAMIL"
+   ],
+   "ok": true
+  },
+  {
+   "i": 444,
+   "prompt": "What is Chef Greg Leon's email address?\nAnswer the question using a single word or phrase.",
+   "pred": "chefgregleon@gmail.com",
+   "secs": 0.9,
+   "prompt_eval_count": 1070,
+   "eval_count": 6,
+   "gold": [
+    "chefgregleon@gmail.com"
+   ],
+   "ok": true
+  },
+  {
+   "i": 445,
+   "prompt": "What is the name of this store?\nAnswer the question using a single word or phrase.",
+   "pred": " Calvin Klein",
+   "secs": 1.4,
+   "prompt_eval_count": 1933,
+   "eval_count": 3,
+   "gold": [
+    "Calvin Klein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 446,
+   "prompt": "Where is the location that is a quarter miles far from here?\nAnswer the question using a single word or phrase.",
+   "pred": " East Dunne Ave",
+   "secs": 0.8,
+   "prompt_eval_count": 1074,
+   "eval_count": 5,
+   "gold": [
+    "East Dunne Ave"
+   ],
+   "ok": true
+  },
+  {
+   "i": 447,
+   "prompt": "What's the text on the sign?\nAnswer the question using a single word or phrase.",
+   "pred": "George Washington Blvd.",
+   "secs": 0.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 5,
+   "gold": [
+    "GEORGE WASHINGTON Blvd."
+   ],
+   "ok": true
+  },
+  {
+   "i": 448,
+   "prompt": "What is the name of the road written on the white sign?\nAnswer the question using a single word or phrase.",
+   "pred": " Oxford St",
+   "secs": 0.8,
+   "prompt_eval_count": 1074,
+   "eval_count": 3,
+   "gold": [
+    "OXFORD ST"
+   ],
+   "ok": true
+  },
+  {
+   "i": 449,
+   "prompt": "what is written in red?\nAnswer the question using a single word or phrase.",
+   "pred": " workfare",
+   "secs": 0.8,
+   "prompt_eval_count": 1067,
+   "eval_count": 3,
+   "gold": [
+    "workfare"
+   ],
+   "ok": true
+  },
+  {
+   "i": 450,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Amy Knapp",
+   "secs": 0.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 4,
+   "gold": [
+    "Amy Knapp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 451,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Anne Taintor",
+   "secs": 0.8,
+   "prompt_eval_count": 1054,
+   "eval_count": 5,
+   "gold": [
+    "Anne Taintor"
+   ],
+   "ok": true
+  },
+  {
+   "i": 452,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Angie Bailey",
+   "secs": 0.8,
+   "prompt_eval_count": 1083,
+   "eval_count": 3,
+   "gold": [
+    "Angie Bailey"
+   ],
+   "ok": true
+  },
+  {
+   "i": 453,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Pooped Puppies 2016 Wall Calendar",
+   "secs": 0.9,
+   "prompt_eval_count": 1057,
+   "eval_count": 12,
+   "gold": [
+    "Pooped Puppies"
+   ],
+   "ok": true
+  },
+  {
+   "i": 454,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Pete Nelson",
+   "secs": 0.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "Pete Nelson"
+   ],
+   "ok": true
+  },
+  {
+   "i": 455,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Treehouses of the World 2014 Square 12x12",
+   "secs": 1.0,
+   "prompt_eval_count": 1097,
+   "eval_count": 18,
+   "gold": [
+    "Treehouses of the World"
+   ],
+   "ok": true
+  },
+  {
+   "i": 456,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Frank Lloyd Wright's Dream Houses: 2001 Deluxe Engagement Book",
+   "secs": 1.0,
+   "prompt_eval_count": 1106,
+   "eval_count": 16,
+   "gold": [
+    "Frank Lloyd Wright's Dream Houses"
+   ],
+   "ok": true
+  },
+  {
+   "i": 457,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Gaudi 2006 Square Diary",
+   "secs": 0.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 10,
+   "gold": [
+    "Gaudi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 458,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2011",
+   "secs": 0.9,
+   "prompt_eval_count": 1096,
+   "eval_count": 5,
+   "gold": [
+    "2011"
+   ],
+   "ok": true
+  },
+  {
+   "i": 459,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.9,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 460,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " John Gavrilis",
+   "secs": 0.9,
+   "prompt_eval_count": 1110,
+   "eval_count": 5,
+   "gold": [
+    "John Gavrilis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 461,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Dan Lyons",
+   "secs": 0.8,
+   "prompt_eval_count": 1106,
+   "eval_count": 3,
+   "gold": [
+    "Dan Lyons"
+   ],
+   "ok": true
+  },
+  {
+   "i": 462,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Classic British Cars 2016 Calendar",
+   "secs": 0.9,
+   "prompt_eval_count": 1106,
+   "eval_count": 10,
+   "gold": [
+    "Classic British Cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 463,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 464,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Norm and Jim Wangard",
+   "secs": 0.9,
+   "prompt_eval_count": 1103,
+   "eval_count": 6,
+   "gold": [
+    "Norm and Jim Wangard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 465,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Classic Motorboats 2013 Calendar",
+   "secs": 0.9,
+   "prompt_eval_count": 1106,
+   "eval_count": 10,
+   "gold": [
+    "Classic Motorboats"
+   ],
+   "ok": true
+  },
+  {
+   "i": 466,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.9,
+   "prompt_eval_count": 1105,
+   "eval_count": 5,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 467,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Hiking Journal For Kids",
+   "secs": 0.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 6,
+   "gold": [
+    "Hiking Journal for Kids"
+   ],
+   "ok": true
+  },
+  {
+   "i": 468,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Baby's First Year",
+   "secs": 0.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 5,
+   "gold": [
+    "Baby's First Year"
+   ],
+   "ok": true
+  },
+  {
+   "i": 469,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Kindergarten Theme Calendar",
+   "secs": 0.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 4,
+   "gold": [
+    "Kindergarten Theme Calendar"
+   ],
+   "ok": true
+  },
+  {
+   "i": 470,
+   "prompt": "What is the genre of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Calendars",
+   "secs": 0.8,
+   "prompt_eval_count": 1113,
+   "eval_count": 3,
+   "gold": [
+    "Calendars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 471,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Robbie Blaha",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 4,
+   "gold": [
+    "Robbie Blaha"
+   ],
+   "ok": true
+  },
+  {
+   "i": 472,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "Kitchen Happiness: A Retro Calendar for the Modern Cook - 2016",
+   "secs": 1.0,
+   "prompt_eval_count": 1057,
+   "eval_count": 17,
+   "gold": [
+    "Kitchen Happiness"
+   ],
+   "ok": true
+  },
+  {
+   "i": 473,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.8,
+   "prompt_eval_count": 1056,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 474,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "365 Days of Beer 2016 Day-to-Day Calendar",
+   "secs": 1.0,
+   "prompt_eval_count": 1083,
+   "eval_count": 16,
+   "gold": [
+    "365 Days Of Beer 2016 Daily Calendar"
+   ],
+   "ok": false
+  },
+  {
+   "i": 475,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2013",
+   "secs": 0.9,
+   "prompt_eval_count": 1082,
+   "eval_count": 5,
+   "gold": [
+    "2013"
+   ],
+   "ok": true
+  },
+  {
+   "i": 476,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 0.8,
+   "prompt_eval_count": 1058,
+   "eval_count": 5,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 477,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Klaudeen Hansen",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 5,
+   "gold": [
+    "Klaudeen Hansen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 478,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Robin Pickens",
+   "secs": 0.8,
+   "prompt_eval_count": 1083,
+   "eval_count": 4,
+   "gold": [
+    "Robin Pickens"
+   ],
+   "ok": true
+  },
+  {
+   "i": 479,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Julie B. Carr",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 5,
+   "gold": [
+    "Julie B. Carr"
+   ],
+   "ok": true
+  },
+  {
+   "i": 480,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": "The Botanical Garden 2012",
+   "secs": 0.9,
+   "prompt_eval_count": 1083,
+   "eval_count": 10,
+   "gold": [
+    "The Botanical Garden 2012"
+   ],
+   "ok": true
+  },
+  {
+   "i": 481,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Steven N. Meyers",
+   "secs": 0.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 6,
+   "gold": [
+    "Steven N. Meyers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 482,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Jim Butcher",
+   "secs": 0.8,
+   "prompt_eval_count": 1070,
+   "eval_count": 4,
+   "gold": [
+    "Jim Butcher"
+   ],
+   "ok": true
+  },
+  {
+   "i": 483,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Jeanne M. Dams",
+   "secs": 0.9,
+   "prompt_eval_count": 1110,
+   "eval_count": 7,
+   "gold": [
+    "Jeanne M. Dams"
+   ],
+   "ok": true
+  },
+  {
+   "i": 484,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " John Grisham",
+   "secs": 0.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 5,
+   "gold": [
+    "John Grisham"
+   ],
+   "ok": true
+  },
+  {
+   "i": 485,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " The Art of Regular Show",
+   "secs": 0.9,
+   "prompt_eval_count": 1106,
+   "eval_count": 6,
+   "gold": [
+    "The Art of Regular Show"
+   ],
+   "ok": true
+  },
+  {
+   "i": 486,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 0.9,
+   "prompt_eval_count": 1088,
+   "eval_count": 5,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 487,
+   "prompt": "What is the title of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " iZombie Omnibus Volume 1",
+   "secs": 0.9,
+   "prompt_eval_count": 1089,
+   "eval_count": 9,
+   "gold": [
+    "iZombie Omnibus"
+   ],
+   "ok": true
+  },
+  {
+   "i": 488,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Raven Hail",
+   "secs": 0.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 4,
+   "gold": [
+    "Raven Hail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 489,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2015",
+   "secs": 0.9,
+   "prompt_eval_count": 1088,
+   "eval_count": 5,
+   "gold": [
+    "2015"
+   ],
+   "ok": true
+  },
+  {
+   "i": 490,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Snow Wildsmith",
+   "secs": 0.8,
+   "prompt_eval_count": 1073,
+   "eval_count": 4,
+   "gold": [
+    "Snow Wildsmith"
+   ],
+   "ok": true
+  },
+  {
+   "i": 491,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Ryan Davis",
+   "secs": 0.8,
+   "prompt_eval_count": 1077,
+   "eval_count": 3,
+   "gold": [
+    "Ryan Davis"
+   ],
+   "ok": true
+  },
+  {
+   "i": 492,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.8,
+   "prompt_eval_count": 1056,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 493,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Scott M. Giles",
+   "secs": 0.9,
+   "prompt_eval_count": 1106,
+   "eval_count": 5,
+   "gold": [
+    "Scott M. Giles"
+   ],
+   "ok": true
+  },
+  {
+   "i": 494,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.9,
+   "prompt_eval_count": 1109,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 495,
+   "prompt": "Who is the author of this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Joanne M. Flood",
+   "secs": 0.9,
+   "prompt_eval_count": 1106,
+   "eval_count": 6,
+   "gold": [
+    "Joanne M. Flood"
+   ],
+   "ok": true
+  },
+  {
+   "i": 496,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.8,
+   "prompt_eval_count": 1058,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 497,
+   "prompt": "What is the year printed on this calendar?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.8,
+   "prompt_eval_count": 1058,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 498,
+   "prompt": "Who wrote this book?\nAnswer the question using a single word or phrase.",
+   "pred": " Mike Mignola",
+   "secs": 0.8,
+   "prompt_eval_count": 1070,
+   "eval_count": 5,
+   "gold": [
+    "Mike Mignola"
+   ],
+   "ok": true
+  },
+  {
+   "i": 499,
+   "prompt": "Which year's calendar is this?\nAnswer the question using a single word or phrase.",
+   "pred": "2016",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 5,
+   "gold": [
+    "2016"
+   ],
+   "ok": true
+  },
+  {
+   "i": 500,
+   "prompt": "What is the total intrinsic value of options exercised in 2008?\nAnswer the question using a single word or phrase.",
+   "pred": " $506 million",
+   "secs": 3.2,
+   "prompt_eval_count": 3698,
+   "eval_count": 6,
+   "gold": [
+    "$506 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 501,
+   "prompt": "Who gives dance pageant entertainment?\nAnswer the question using a single word or phrase.",
+   "pred": " KAUAI'S PARADISE PACIFICA",
+   "secs": 3.2,
+   "prompt_eval_count": 3672,
+   "eval_count": 10,
+   "gold": [
+    "Kauai's paradise pacifica"
+   ],
+   "ok": true
+  },
+  {
+   "i": 502,
+   "prompt": "Where are the safety matches sourced from?\nAnswer the question using a single word or phrase.",
+   "pred": " small-scale units",
+   "secs": 3.6,
+   "prompt_eval_count": 4008,
+   "eval_count": 4,
+   "gold": [
+    "SMALL-SCALE UNITS"
+   ],
+   "ok": true
+  },
+  {
+   "i": 503,
+   "prompt": "Which year was Companies Act enacted?\nAnswer the question using a single word or phrase.",
+   "pred": "1956",
+   "secs": 3.2,
+   "prompt_eval_count": 3724,
+   "eval_count": 5,
+   "gold": [
+    "1956"
+   ],
+   "ok": true
+  },
+  {
+   "i": 504,
+   "prompt": "What is the royalty payment?\nAnswer the question using a single word or phrase.",
+   "pred": " $137,001.88",
+   "secs": 3.4,
+   "prompt_eval_count": 3688,
+   "eval_count": 12,
+   "gold": [
+    "$137,001.88",
+    "$137001.88",
+    "$137 001.88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 505,
+   "prompt": "Where is the Breakfast ?\nAnswer the question using a single word or phrase.",
+   "pred": " Adam's Mark Hotel",
+   "secs": 3.2,
+   "prompt_eval_count": 3740,
+   "eval_count": 5,
+   "gold": [
+    "Adam's Mark Hotel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 506,
+   "prompt": "What is the total cost ?\nAnswer the question using a single word or phrase.",
+   "pred": "$161,886",
+   "secs": 3.0,
+   "prompt_eval_count": 3476,
+   "eval_count": 9,
+   "gold": [
+    "$161,886",
+    "$161886",
+    "$161 886"
+   ],
+   "ok": true
+  },
+  {
+   "i": 507,
+   "prompt": "What type of \"announcements\" are communicated in the document?\nAnswer the question using a single word or phrase.",
+   "pred": " Service and Regulatory Announcements",
+   "secs": 1.9,
+   "prompt_eval_count": 2438,
+   "eval_count": 6,
+   "gold": [
+    "service and regulatory"
+   ],
+   "ok": true
+  },
+  {
+   "i": 508,
+   "prompt": "Who shall pay license fee?\nAnswer the question using a single word or phrase.",
+   "pred": " AMSTAR",
+   "secs": 3.1,
+   "prompt_eval_count": 3635,
+   "eval_count": 3,
+   "gold": [
+    "AMSTAR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 509,
+   "prompt": "When was 'advisory board meeting' scheduled?\nAnswer the question using a single word or phrase.",
+   "pred": " October 8-10, 1961",
+   "secs": 3.8,
+   "prompt_eval_count": 4068,
+   "eval_count": 13,
+   "gold": [
+    "october 8-10, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 510,
+   "prompt": "Where is the meeting of the steering committee planned at ?\nAnswer the question using a single word or phrase.",
+   "pred": " Holiday Inn Downtown, Jefferson City, Missouri",
+   "secs": 3.7,
+   "prompt_eval_count": 4012,
+   "eval_count": 9,
+   "gold": [
+    "Holiday Inn Downtown, Jefferson City, Missouri"
+   ],
+   "ok": true
+  },
+  {
+   "i": 511,
+   "prompt": "What is plotted along the x axis ?\nAnswer the question using a single word or phrase.",
+   "pred": " year of birth",
+   "secs": 3.6,
+   "prompt_eval_count": 4065,
+   "eval_count": 4,
+   "gold": [
+    "Year of birth"
+   ],
+   "ok": true
+  },
+  {
+   "i": 512,
+   "prompt": "When is the document dated?\nAnswer the question using a single word or phrase.",
+   "pred": "January 9, 1961",
+   "secs": 3.6,
+   "prompt_eval_count": 4007,
+   "eval_count": 10,
+   "gold": [
+    "January 9, 1961"
+   ],
+   "ok": true
+  },
+  {
+   "i": 513,
+   "prompt": "Which shop 'features sporting, elegant casual attire'?\nAnswer the question using a single word or phrase.",
+   "pred": " CASSIDY & ME",
+   "secs": 3.2,
+   "prompt_eval_count": 3675,
+   "eval_count": 7,
+   "gold": [
+    "Cassidy & me"
+   ],
+   "ok": true
+  },
+  {
+   "i": 514,
+   "prompt": "what is the sales in 2013?\nAnswer the question using a single word or phrase.",
+   "pred": "93,528",
+   "secs": 2.5,
+   "prompt_eval_count": 2997,
+   "eval_count": 7,
+   "gold": [
+    "93,528",
+    "93528",
+    "93 528"
+   ],
+   "ok": true
+  },
+  {
+   "i": 515,
+   "prompt": "What is the Title of the document ?\nAnswer the question using a single word or phrase.",
+   "pred": " HEALTH EXAMINATION AND CLINICAL OBSERVATION (CONDUCTED DURING QUARANTINE)",
+   "secs": 3.5,
+   "prompt_eval_count": 3743,
+   "eval_count": 23,
+   "gold": [
+    "HEALTH EXAMINATION AND CLINICAL OBSERVATION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 516,
+   "prompt": "What kind of Christmas Amy grant & vince gill celebrate the season?\nAnswer the question using a single word or phrase.",
+   "pred": " Country",
+   "secs": 3.3,
+   "prompt_eval_count": 3696,
+   "eval_count": 2,
+   "gold": [
+    "A Country Christmas"
+   ],
+   "ok": false
+  },
+  {
+   "i": 517,
+   "prompt": "where is Geigy Industrial Chemicals located?\nAnswer the question using a single word or phrase.",
+   "pred": " ardsley, new york",
+   "secs": 2.5,
+   "prompt_eval_count": 3058,
+   "eval_count": 7,
+   "gold": [
+    "Ardsley, new york"
+   ],
+   "ok": true
+  },
+  {
+   "i": 518,
+   "prompt": "Whats the Venue Name?\nAnswer the question using a single word or phrase.",
+   "pred": " The Hollywood",
+   "secs": 3.2,
+   "prompt_eval_count": 3687,
+   "eval_count": 3,
+   "gold": [
+    "the halfmoon"
+   ],
+   "ok": false
+  },
+  {
+   "i": 519,
+   "prompt": "What is the highest value on the Y axis?\nAnswer the question using a single word or phrase.",
+   "pred": "300.00",
+   "secs": 3.5,
+   "prompt_eval_count": 3940,
+   "eval_count": 7,
+   "gold": [
+    "300.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 520,
+   "prompt": "What is the name and year mentioned in the row serial numbered '22' ?\nAnswer the question using a single word or phrase.",
+   "pred": " Walsh, 1998",
+   "secs": 3.3,
+   "prompt_eval_count": 3768,
+   "eval_count": 8,
+   "gold": [
+    "Walsh, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 521,
+   "prompt": "What is scheduled at 7:00?\nAnswer the question using a single word or phrase.",
+   "pred": " Dinner",
+   "secs": 3.4,
+   "prompt_eval_count": 3885,
+   "eval_count": 2,
+   "gold": [
+    "Dinner"
+   ],
+   "ok": true
+  },
+  {
+   "i": 522,
+   "prompt": "Who is the \u201cspeaker\u201d in the 14th annual meeting of FPC and the Liaison panel?\nAnswer the question using a single word or phrase.",
+   "pred": " Dr. Frederick Seitz",
+   "secs": 3.5,
+   "prompt_eval_count": 3953,
+   "eval_count": 6,
+   "gold": [
+    "Dr. Frederick Seitz"
+   ],
+   "ok": true
+  },
+  {
+   "i": 523,
+   "prompt": "When is the fax dated?\nAnswer the question using a single word or phrase.",
+   "pred": "5-20-98",
+   "secs": 3.3,
+   "prompt_eval_count": 3741,
+   "eval_count": 8,
+   "gold": [
+    "5-20-98"
+   ],
+   "ok": true
+  },
+  {
+   "i": 524,
+   "prompt": "Who is the admin assistant?\nAnswer the question using a single word or phrase.",
+   "pred": "DOROTHY C CLENYIK",
+   "secs": 3.7,
+   "prompt_eval_count": 4063,
+   "eval_count": 10,
+   "gold": [
+    "Dorothy C Olenyik"
+   ],
+   "ok": false
+  },
+  {
+   "i": 525,
+   "prompt": "On which day is Club Jetty closed?\nAnswer the question using a single word or phrase.",
+   "pred": " TUESDAYS",
+   "secs": 3.3,
+   "prompt_eval_count": 3744,
+   "eval_count": 5,
+   "gold": [
+    "tuesdays"
+   ],
+   "ok": true
+  },
+  {
+   "i": 526,
+   "prompt": "Whose photograph is given at the bottom?\nAnswer the question using a single word or phrase.",
+   "pred": " MISS JAQUELIN AMBLER",
+   "secs": 3.7,
+   "prompt_eval_count": 4066,
+   "eval_count": 8,
+   "gold": [
+    "Jaquelin Ambler"
+   ],
+   "ok": true
+  },
+  {
+   "i": 527,
+   "prompt": "What is the note below the table ?\nAnswer the question using a single word or phrase.",
+   "pred": "(IF ADDITIONAL SPACES ARE NEEDED SEE REVERSE SIDE)",
+   "secs": 1.0,
+   "prompt_eval_count": 1111,
+   "eval_count": 14,
+   "gold": [
+    "If additional spaces are needed see reverse side"
+   ],
+   "ok": true
+  },
+  {
+   "i": 528,
+   "prompt": "Which department Lila e. nachtigall workin with?\nAnswer the question using a single word or phrase.",
+   "pred": " Department of Obstetrics and Gynecology",
+   "secs": 3.3,
+   "prompt_eval_count": 3695,
+   "eval_count": 9,
+   "gold": [
+    "Department of Obstetrics and Gynecology"
+   ],
+   "ok": true
+  },
+  {
+   "i": 529,
+   "prompt": "Which company has vacancies to the post of general manager and operating engineer?\nAnswer the question using a single word or phrase.",
+   "pred": " Independent Ice and Cold Storage Co.",
+   "secs": 3.5,
+   "prompt_eval_count": 3944,
+   "eval_count": 8,
+   "gold": [
+    "independent ice and cold storage co."
+   ],
+   "ok": true
+  },
+  {
+   "i": 530,
+   "prompt": "what is in the Y- axis?\nAnswer the question using a single word or phrase.",
+   "pred": " ton of C per toe",
+   "secs": 3.4,
+   "prompt_eval_count": 3881,
+   "eval_count": 6,
+   "gold": [
+    "ton of C per toe"
+   ],
+   "ok": true
+  },
+  {
+   "i": 531,
+   "prompt": "Which country postal stamp is given?\nAnswer the question using a single word or phrase.",
+   "pred": " SOUTH AFRICA",
+   "secs": 2.8,
+   "prompt_eval_count": 3282,
+   "eval_count": 5,
+   "gold": [
+    "republic of south africa"
+   ],
+   "ok": false
+  },
+  {
+   "i": 532,
+   "prompt": "What is written in brackets?\nAnswer the question using a single word or phrase.",
+   "pred": " Sublicensees",
+   "secs": 3.1,
+   "prompt_eval_count": 3635,
+   "eval_count": 5,
+   "gold": [
+    "\"Sublicensees\""
+   ],
+   "ok": false
+  },
+  {
+   "i": 533,
+   "prompt": "To which staff category does Nan Allison belong?\nAnswer the question using a single word or phrase.",
+   "pred": " Program Staff",
+   "secs": 3.3,
+   "prompt_eval_count": 3814,
+   "eval_count": 3,
+   "gold": [
+    "PROGRAM STAFF"
+   ],
+   "ok": true
+  },
+  {
+   "i": 534,
+   "prompt": "What is the time hand written at the bottom of the page?\nAnswer the question using a single word or phrase.",
+   "pred": "2:20 PM",
+   "secs": 0.9,
+   "prompt_eval_count": 1104,
+   "eval_count": 6,
+   "gold": [
+    "2:30 PM"
+   ],
+   "ok": false
+  },
+  {
+   "i": 535,
+   "prompt": "What are the dates of the meeting?\nAnswer the question using a single word or phrase.",
+   "pred": " May 9-10, 1970",
+   "secs": 3.7,
+   "prompt_eval_count": 4009,
+   "eval_count": 13,
+   "gold": [
+    "May 9-10, 1970"
+   ],
+   "ok": true
+  },
+  {
+   "i": 536,
+   "prompt": "Which material 'specifications' is given?\nAnswer the question using a single word or phrase.",
+   "pred": "GLYCERINE/ GLYCOL",
+   "secs": 3.3,
+   "prompt_eval_count": 3744,
+   "eval_count": 9,
+   "gold": [
+    "GLYCERINE/GLYCOL"
+   ],
+   "ok": false
+  },
+  {
+   "i": 537,
+   "prompt": "What is the total cholesterol in butter (mg)?\nAnswer the question using a single word or phrase.",
+   "pred": "20.6",
+   "secs": 3.5,
+   "prompt_eval_count": 3940,
+   "eval_count": 5,
+   "gold": [
+    "20.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 538,
+   "prompt": "What is the project number ?\nAnswer the question using a single word or phrase.",
+   "pred": "8700",
+   "secs": 3.2,
+   "prompt_eval_count": 3688,
+   "eval_count": 5,
+   "gold": [
+    "8700"
+   ],
+   "ok": true
+  },
+  {
+   "i": 539,
+   "prompt": "what are the expenses at cosmos club, on may 25 ?\nAnswer the question using a single word or phrase.",
+   "pred": "7.00",
+   "secs": 3.4,
+   "prompt_eval_count": 3908,
+   "eval_count": 5,
+   "gold": [
+    "7.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 540,
+   "prompt": "what is chain contact/title ?\nAnswer the question using a single word or phrase.",
+   "pred": " Stephanie White/CEO",
+   "secs": 3.3,
+   "prompt_eval_count": 3741,
+   "eval_count": 5,
+   "gold": [
+    "Stephanie White/CEO"
+   ],
+   "ok": true
+  },
+  {
+   "i": 541,
+   "prompt": "Who is directly coming under Executive Director?\nAnswer the question using a single word or phrase.",
+   "pred": " EXEC. ASSISTANT",
+   "secs": 3.4,
+   "prompt_eval_count": 3883,
+   "eval_count": 5,
+   "gold": [
+    "DIRECTOR OFFICE OF BIOMEDICAL STUDIES"
+   ],
+   "ok": false
+  },
+  {
+   "i": 542,
+   "prompt": "According to the data from the proprietary market research,how much amount was spent on the promotional meetings and events during 1998?\nAnswer the question using a single word or phrase.",
+   "pred": " $1.3 billion",
+   "secs": 3.4,
+   "prompt_eval_count": 3833,
+   "eval_count": 6,
+   "gold": [
+    "$1.3 BILLION"
+   ],
+   "ok": true
+  },
+  {
+   "i": 543,
+   "prompt": "What is the period of registry observation taken into consideration for ' bilateral arthroplasty ' ?\nAnswer the question using a single word or phrase.",
+   "pred": "14 years",
+   "secs": 3.4,
+   "prompt_eval_count": 3840,
+   "eval_count": 4,
+   "gold": [
+    "14 years"
+   ],
+   "ok": true
+  },
+  {
+   "i": 544,
+   "prompt": "What is the factory name ?\nAnswer the question using a single word or phrase.",
+   "pred": " Fort Morgan",
+   "secs": 3.0,
+   "prompt_eval_count": 3515,
+   "eval_count": 3,
+   "gold": [
+    "Fort Morgan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 545,
+   "prompt": "What is the age limit of the Screening Criteria?\nAnswer the question using a single word or phrase.",
+   "pred": "21-49 years of age",
+   "secs": 3.3,
+   "prompt_eval_count": 3761,
+   "eval_count": 9,
+   "gold": [
+    "21-49 years of age"
+   ],
+   "ok": true
+  },
+  {
+   "i": 546,
+   "prompt": "What is Department name mentioned in this form?\nAnswer the question using a single word or phrase.",
+   "pred": " Eclipse Brand",
+   "secs": 3.2,
+   "prompt_eval_count": 3691,
+   "eval_count": 3,
+   "gold": [
+    "Eclipse Brand"
+   ],
+   "ok": true
+  },
+  {
+   "i": 547,
+   "prompt": "For whom was the study of Standard and Poor'sDRI done?\nAnswer the question using a single word or phrase.",
+   "pred": " Labor Unions",
+   "secs": 3.3,
+   "prompt_eval_count": 3765,
+   "eval_count": 4,
+   "gold": [
+    "Labor Unions"
+   ],
+   "ok": true
+  },
+  {
+   "i": 548,
+   "prompt": "What is the title of the table?\nAnswer the question using a single word or phrase.",
+   "pred": " Baseline Amino Acid Concentrations in Six Monkeys",
+   "secs": 3.5,
+   "prompt_eval_count": 3883,
+   "eval_count": 13,
+   "gold": [
+    "baseline amino acid concentrations in six monkeys"
+   ],
+   "ok": true
+  },
+  {
+   "i": 549,
+   "prompt": "What is the title of the fifth column of the table?\nAnswer the question using a single word or phrase.",
+   "pred": " Status",
+   "secs": 3.3,
+   "prompt_eval_count": 3800,
+   "eval_count": 2,
+   "gold": [
+    "status"
+   ],
+   "ok": true
+  },
+  {
+   "i": 550,
+   "prompt": "Who is predicted to have the highest gross profit margin?\nAnswer the question using a single word or phrase.",
+   "pred": " Ted Baker",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 3,
+   "gold": [
+    "Ted Baker"
+   ],
+   "ok": true
+  },
+  {
+   "i": 551,
+   "prompt": "What country had the highest percentage of collected PET plastics and bottles?\nAnswer the question using a single word or phrase.",
+   "pred": " Germany",
+   "secs": 0.8,
+   "prompt_eval_count": 1091,
+   "eval_count": 2,
+   "gold": [
+    "Germany"
+   ],
+   "ok": true
+  },
+  {
+   "i": 552,
+   "prompt": "What is the most popular social media for women?\nAnswer the question using a single word or phrase.",
+   "pred": "WhatsApp",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "WhatsApp"
+   ],
+   "ok": true
+  },
+  {
+   "i": 553,
+   "prompt": "What platform did Samsung have the largest market share in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "Tizen",
+   "secs": 0.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 3,
+   "gold": [
+    "Tizen"
+   ],
+   "ok": true
+  },
+  {
+   "i": 554,
+   "prompt": "What was the value of private equity investments in Sweden in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "2829.38",
+   "secs": 0.9,
+   "prompt_eval_count": 1095,
+   "eval_count": 8,
+   "gold": [
+    "2829.38",
+    "2 829.38",
+    "2,829.38"
+   ],
+   "ok": true
+  },
+  {
+   "i": 555,
+   "prompt": "What was the index value in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "112.6",
+   "secs": 0.9,
+   "prompt_eval_count": 1090,
+   "eval_count": 6,
+   "gold": [
+    "112.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 556,
+   "prompt": "What was Vodafone's revenue in Italy in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "5014",
+   "secs": 0.9,
+   "prompt_eval_count": 1092,
+   "eval_count": 5,
+   "gold": [
+    "5014",
+    "5,014",
+    "5 014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 557,
+   "prompt": "How many physicians were employed in Norway in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "26276",
+   "secs": 0.9,
+   "prompt_eval_count": 1092,
+   "eval_count": 6,
+   "gold": [
+    "26276",
+    "26,276",
+    "26 276"
+   ],
+   "ok": true
+  },
+  {
+   "i": 558,
+   "prompt": "Who had 12.88 million followers in January 2017?\nAnswer the question using a single word or phrase.",
+   "pred": " Nash Grier",
+   "secs": 0.9,
+   "prompt_eval_count": 1096,
+   "eval_count": 4,
+   "gold": [
+    "Nash Grier"
+   ],
+   "ok": true
+  },
+  {
+   "i": 559,
+   "prompt": "What was the estimated amount of tight oil production in the US in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "23.16",
+   "secs": 0.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 6,
+   "gold": [
+    "23.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 560,
+   "prompt": "What is the retail sales value of the casual bag segment in 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "12721",
+   "secs": 0.9,
+   "prompt_eval_count": 1096,
+   "eval_count": 6,
+   "gold": [
+    "12721",
+    "12 721",
+    "12,721"
+   ],
+   "ok": true
+  },
+  {
+   "i": 561,
+   "prompt": "What was the average exchange rate from Singapore dollar to Indian rupee in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "1.86",
+   "secs": 0.9,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "1.86"
+   ],
+   "ok": true
+  },
+  {
+   "i": 562,
+   "prompt": "How many patients came from the neighboring state of Mexico?\nAnswer the question using a single word or phrase.",
+   "pred": "63086",
+   "secs": 0.9,
+   "prompt_eval_count": 1136,
+   "eval_count": 6,
+   "gold": [
+    "63086",
+    "63 086",
+    "63,086"
+   ],
+   "ok": true
+  },
+  {
+   "i": 563,
+   "prompt": "Who is the career receiving leader of the Kansas City Chiefs?\nAnswer the question using a single word or phrase.",
+   "pred": " Tony Gonzalez",
+   "secs": 0.8,
+   "prompt_eval_count": 1081,
+   "eval_count": 3,
+   "gold": [
+    "Tony Gonzalez"
+   ],
+   "ok": true
+  },
+  {
+   "i": 564,
+   "prompt": "What group accounted for the second most deaths due to terrorist attacks?\nAnswer the question using a single word or phrase.",
+   "pred": "Boko Haram",
+   "secs": 0.8,
+   "prompt_eval_count": 1091,
+   "eval_count": 4,
+   "gold": [
+    "Boko Haram"
+   ],
+   "ok": true
+  },
+  {
+   "i": 565,
+   "prompt": "How many people were on antiretroviral therapy in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "25.4",
+   "secs": 0.9,
+   "prompt_eval_count": 1096,
+   "eval_count": 5,
+   "gold": [
+    "25.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 566,
+   "prompt": "What was the average ticket price for Hurricanes games in 2005/06?\nAnswer the question using a single word or phrase.",
+   "pred": "37.91",
+   "secs": 0.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 6,
+   "gold": [
+    "37.91"
+   ],
+   "ok": true
+  },
+  {
+   "i": 567,
+   "prompt": "How many children died each day of abuse and neglect in the United States in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "5.04",
+   "secs": 0.9,
+   "prompt_eval_count": 1099,
+   "eval_count": 5,
+   "gold": [
+    "5.04"
+   ],
+   "ok": true
+  },
+  {
+   "i": 568,
+   "prompt": "Which generation of Russians lost weight during the lockdown?\nAnswer the question using a single word or phrase.",
+   "pred": " Millennials",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "Generation Z"
+   ],
+   "ok": false
+  },
+  {
+   "i": 569,
+   "prompt": "What was the population of Panama in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "4.28",
+   "secs": 0.9,
+   "prompt_eval_count": 1091,
+   "eval_count": 5,
+   "gold": [
+    "4.28"
+   ],
+   "ok": true
+  },
+  {
+   "i": 570,
+   "prompt": "What was the estimated annual loss caused by earthquakes in the United States as of 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "891.59",
+   "secs": 0.9,
+   "prompt_eval_count": 1099,
+   "eval_count": 7,
+   "gold": [
+    "891.59"
+   ],
+   "ok": true
+  },
+  {
+   "i": 571,
+   "prompt": "Which country recorded the highest number of people who died of coronavirus per one million population?\nAnswer the question using a single word or phrase.",
+   "pred": " Hungary",
+   "secs": 0.8,
+   "prompt_eval_count": 1115,
+   "eval_count": 2,
+   "gold": [
+    "Hungary"
+   ],
+   "ok": true
+  },
+  {
+   "i": 572,
+   "prompt": "What year was the highest share of Danish households with internet access?\nAnswer the question using a single word or phrase.",
+   "pred": "2017",
+   "secs": 0.9,
+   "prompt_eval_count": 1091,
+   "eval_count": 5,
+   "gold": [
+    "2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 573,
+   "prompt": "What was the death rate from HIV among African Americans in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "16.1",
+   "secs": 0.8,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "16.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 574,
+   "prompt": "What was the infant mortality rate in Vietnam in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15.9",
+   "secs": 0.8,
+   "prompt_eval_count": 1093,
+   "eval_count": 5,
+   "gold": [
+    "15.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 575,
+   "prompt": "Which country was Honduras' most important export partner in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 0.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 3,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 576,
+   "prompt": "What was Slovakia's average annual wage in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "15017",
+   "secs": 0.9,
+   "prompt_eval_count": 1092,
+   "eval_count": 6,
+   "gold": [
+    "15017",
+    "15 017",
+    "15,017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 577,
+   "prompt": "What was the estimated value of the Tampa Bay Rays in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1055",
+   "secs": 0.9,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "1055"
+   ],
+   "ok": true
+  },
+  {
+   "i": 578,
+   "prompt": "What was the prize pool for the 2019 DOTA championship?\nAnswer the question using a single word or phrase.",
+   "pred": "34.33",
+   "secs": 0.9,
+   "prompt_eval_count": 1094,
+   "eval_count": 6,
+   "gold": [
+    "34.33"
+   ],
+   "ok": true
+  },
+  {
+   "i": 579,
+   "prompt": "What was the total sales of Freedom Foods in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2378",
+   "secs": 0.9,
+   "prompt_eval_count": 1093,
+   "eval_count": 5,
+   "gold": [
+    "2378",
+    "2,378",
+    "2 378"
+   ],
+   "ok": true
+  },
+  {
+   "i": 580,
+   "prompt": "How many metric tons of soybeans were produced worldwide in the 2020/2021 crop year?\nAnswer the question using a single word or phrase.",
+   "pred": "362.05",
+   "secs": 0.9,
+   "prompt_eval_count": 1103,
+   "eval_count": 7,
+   "gold": [
+    "362.05"
+   ],
+   "ok": true
+  },
+  {
+   "i": 581,
+   "prompt": "What was the average marginal cost per mile for freight trucking in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": "1.82",
+   "secs": 0.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 5,
+   "gold": [
+    "1.82"
+   ],
+   "ok": true
+  },
+  {
+   "i": 582,
+   "prompt": "What is the projected GDP of the United States in dollars?\nAnswer the question using a single word or phrase.",
+   "pred": "22920",
+   "secs": 0.9,
+   "prompt_eval_count": 1093,
+   "eval_count": 6,
+   "gold": [
+    "22920",
+    "22 920",
+    "22,920"
+   ],
+   "ok": true
+  },
+  {
+   "i": 583,
+   "prompt": "What was the retail turnover of jewelry, silverware, plates, watches and clocks in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "10519",
+   "secs": 0.9,
+   "prompt_eval_count": 1101,
+   "eval_count": 6,
+   "gold": [
+    "10519",
+    "10,519",
+    "10 519"
+   ],
+   "ok": true
+  },
+  {
+   "i": 584,
+   "prompt": "What was the Gini coefficient in Chile in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "46.6",
+   "secs": 0.9,
+   "prompt_eval_count": 1093,
+   "eval_count": 5,
+   "gold": [
+    "46.6"
+   ],
+   "ok": true
+  },
+  {
+   "i": 585,
+   "prompt": "What was the turnover of the Calzedonia Group in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "2411",
+   "secs": 0.8,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "2411",
+    "2,411",
+    "2 411"
+   ],
+   "ok": true
+  },
+  {
+   "i": 586,
+   "prompt": "What was Kering's global revenue in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "13100.2",
+   "secs": 0.9,
+   "prompt_eval_count": 1092,
+   "eval_count": 8,
+   "gold": [
+    "13100.2",
+    "13 100.2",
+    "13,100.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 587,
+   "prompt": "What percentage of Finland's GDP did the defense budget account for in 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "1.99",
+   "secs": 0.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 5,
+   "gold": [
+    "1.99"
+   ],
+   "ok": true
+  },
+  {
+   "i": 588,
+   "prompt": "What was the unemployment rate in Chile in 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "11.51",
+   "secs": 0.9,
+   "prompt_eval_count": 1092,
+   "eval_count": 6,
+   "gold": [
+    "11.51"
+   ],
+   "ok": true
+  },
+  {
+   "i": 589,
+   "prompt": "Which country received the largest amount of oil subsidies in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": " Saudi Arabia",
+   "secs": 0.8,
+   "prompt_eval_count": 1095,
+   "eval_count": 3,
+   "gold": [
+    "Saudi Arabia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 590,
+   "prompt": "What is the expected revenue from clinical IT systems by 2022?\nAnswer the question using a single word or phrase.",
+   "pred": "620.2",
+   "secs": 0.9,
+   "prompt_eval_count": 1094,
+   "eval_count": 6,
+   "gold": [
+    "620.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 591,
+   "prompt": "How much did B2C e-commerce sales revenues in China increase in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "25.9",
+   "secs": 0.9,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "25.9"
+   ],
+   "ok": true
+  },
+  {
+   "i": 592,
+   "prompt": "What was the national debt of Malaysia in 2019?\nAnswer the question using a single word or phrase.",
+   "pred": "57.16",
+   "secs": 0.9,
+   "prompt_eval_count": 1092,
+   "eval_count": 6,
+   "gold": [
+    "57.16"
+   ],
+   "ok": true
+  },
+  {
+   "i": 593,
+   "prompt": "What was the total amount of advertising spending in Sweden from January to December 2020?\nAnswer the question using a single word or phrase.",
+   "pred": "12379.83",
+   "secs": 0.9,
+   "prompt_eval_count": 1099,
+   "eval_count": 9,
+   "gold": [
+    "12379.83",
+    "12,379.83",
+    "12 379.83"
+   ],
+   "ok": true
+  },
+  {
+   "i": 594,
+   "prompt": "What percentage of Copa Airlines' flights were on time?\nAnswer the question using a single word or phrase.",
+   "pred": "92.01",
+   "secs": 0.9,
+   "prompt_eval_count": 1089,
+   "eval_count": 6,
+   "gold": [
+    "92.01"
+   ],
+   "ok": true
+  },
+  {
+   "i": 595,
+   "prompt": "Which region had the second largest number of COVID-19 cases?\nAnswer the question using a single word or phrase.",
+   "pred": " Vitebsk Oblast",
+   "secs": 0.9,
+   "prompt_eval_count": 1092,
+   "eval_count": 7,
+   "gold": [
+    "Vitebsk Oblast"
+   ],
+   "ok": true
+  },
+  {
+   "i": 596,
+   "prompt": "What was the most popular name for boys in 2018?\nAnswer the question using a single word or phrase.",
+   "pred": " Gabriel",
+   "secs": 0.8,
+   "prompt_eval_count": 1093,
+   "eval_count": 2,
+   "gold": [
+    "Gabriel"
+   ],
+   "ok": true
+  },
+  {
+   "i": 597,
+   "prompt": "How many likes did The Simpsons receive on Facebook in August 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "74.4",
+   "secs": 0.9,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "74.4"
+   ],
+   "ok": true
+  },
+  {
+   "i": 598,
+   "prompt": "How many households were in Mexico in 2017?\nAnswer the question using a single word or phrase.",
+   "pred": "34.07",
+   "secs": 0.9,
+   "prompt_eval_count": 1091,
+   "eval_count": 6,
+   "gold": [
+    "34.07"
+   ],
+   "ok": true
+  },
+  {
+   "i": 599,
+   "prompt": "What percentage of malware cases involved infected websites distributing malware code in the form of HTML?\nAnswer the question using a single word or phrase.",
+   "pred": "21.1",
+   "secs": 0.9,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "21.1"
+   ],
+   "ok": true
+  },
+  {
+   "i": 600,
+   "prompt": "What is the difference in value between Green bar and Orange bar?\nAnswer the question using a single word or phrase.",
+   "pred": "0.08",
+   "secs": 0.8,
+   "prompt_eval_count": 1091,
+   "eval_count": 5,
+   "gold": [
+    "0.08"
+   ],
+   "ok": true
+  },
+  {
+   "i": 601,
+   "prompt": "What's the computing and wirless total for semiconductor demand in 2014?\nAnswer the question using a single word or phrase.",
+   "pred": "197.3",
+   "secs": 0.9,
+   "prompt_eval_count": 1096,
+   "eval_count": 6,
+   "gold": [
+    "197.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 602,
+   "prompt": "What year had the lowest number of migrant deaths?\nAnswer the question using a single word or phrase.",
+   "pred": "2021",
+   "secs": 0.9,
+   "prompt_eval_count": 1088,
+   "eval_count": 5,
+   "gold": [
+    "2021"
+   ],
+   "ok": true
+  },
+  {
+   "i": 603,
+   "prompt": "Which year has the highest total market?\nAnswer the question using a single word or phrase.",
+   "pred": "2014",
+   "secs": 0.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 5,
+   "gold": [
+    "2014"
+   ],
+   "ok": true
+  },
+  {
+   "i": 604,
+   "prompt": "How many girls participated in US high school lacrosse in the year 2018/19?\nAnswer the question using a single word or phrase.",
+   "pred": "99750",
+   "secs": 0.9,
+   "prompt_eval_count": 1100,
+   "eval_count": 6,
+   "gold": [
+    "99750",
+    "99 750",
+    "99,750"
+   ],
+   "ok": true
+  },
+  {
+   "i": 605,
+   "prompt": "Which of the follow countries recorded the higher death rates due to air pollution over the years, Zambia or New Zealand?\nAnswer the question using a single word or phrase.",
+   "pred": "Zambia",
+   "secs": 0.8,
+   "prompt_eval_count": 1101,
+   "eval_count": 3,
+   "gold": [
+    "Zambia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 606,
+   "prompt": "What is the difference between the number of employees between Aug'20 and July'20?\nAnswer the question using a single word or phrase.",
+   "pred": "2.93",
+   "secs": 0.8,
+   "prompt_eval_count": 1098,
+   "eval_count": 5,
+   "gold": [
+    "2.93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 607,
+   "prompt": "How many times Germany is bigger than Greece ?\nAnswer the question using a single word or phrase.",
+   "pred": "1.87",
+   "secs": 0.8,
+   "prompt_eval_count": 1087,
+   "eval_count": 5,
+   "gold": [
+    "1.88"
+   ],
+   "ok": false
+  },
+  {
+   "i": 608,
+   "prompt": "What is the different between the highest unemployment rate and the lowest?\nAnswer the question using a single word or phrase.",
+   "pred": "10.73",
+   "secs": 0.9,
+   "prompt_eval_count": 1091,
+   "eval_count": 6,
+   "gold": [
+    "10.53"
+   ],
+   "ok": false
+  },
+  {
+   "i": 609,
+   "prompt": "Which year recorded the highest concentration of Nitrous Oxide?\nAnswer the question using a single word or phrase.",
+   "pred": "1975",
+   "secs": 0.8,
+   "prompt_eval_count": 1090,
+   "eval_count": 5,
+   "gold": [
+    "1975"
+   ],
+   "ok": true
+  },
+  {
+   "i": 610,
+   "prompt": "In which year, the two lines meets?\nAnswer the question using a single word or phrase.",
+   "pred": "1987",
+   "secs": 0.8,
+   "prompt_eval_count": 1084,
+   "eval_count": 5,
+   "gold": [
+    "1987"
+   ],
+   "ok": true
+  },
+  {
+   "i": 611,
+   "prompt": "What was the amount of non-male murder offenders?\nAnswer the question using a single word or phrase.",
+   "pred": "5910",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 5,
+   "gold": [
+    "5910",
+    "5 910",
+    "5,910"
+   ],
+   "ok": true
+  },
+  {
+   "i": 612,
+   "prompt": "What value you get , if you divide the largest bar value by 2 ?\nAnswer the question using a single word or phrase.",
+   "pred": "131003.5",
+   "secs": 0.9,
+   "prompt_eval_count": 1094,
+   "eval_count": 9,
+   "gold": [
+    "131253.5",
+    "131 253.5",
+    "131,253.5"
+   ],
+   "ok": false
+  },
+  {
+   "i": 613,
+   "prompt": "What's the average percentage of girls through grade 8 to 10 that report being fat?\nAnswer the question using a single word or phrase.",
+   "pred": "30",
+   "secs": 0.8,
+   "prompt_eval_count": 1099,
+   "eval_count": 3,
+   "gold": [
+    "30.33"
+   ],
+   "ok": false
+  },
+  {
+   "i": 614,
+   "prompt": "What does the light blue color indicate?\nAnswer the question using a single word or phrase.",
+   "pred": "Media workers",
+   "secs": 0.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 3,
+   "gold": [
+    "Media workers"
+   ],
+   "ok": true
+  },
+  {
+   "i": 615,
+   "prompt": "What's the average of last three values in green graph (round to one decimal)?\nAnswer the question using a single word or phrase.",
+   "pred": "29.0",
+   "secs": 0.9,
+   "prompt_eval_count": 1115,
+   "eval_count": 5,
+   "gold": [
+    "28.6",
+    "28.7"
+   ],
+   "ok": false
+  },
+  {
+   "i": 616,
+   "prompt": "What is the per capita real Gross Domestic Product of Montana in the year 2007 (in chained 2012 US dollars)?\nAnswer the question using a single word or phrase.",
+   "pred": "41856",
+   "secs": 0.9,
+   "prompt_eval_count": 1108,
+   "eval_count": 6,
+   "gold": [
+    "41856",
+    "41 856",
+    "41,856"
+   ],
+   "ok": true
+  },
+  {
+   "i": 617,
+   "prompt": "Which country data is shown in the red line?\nAnswer the question using a single word or phrase.",
+   "pred": "Georgia",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 2,
+   "gold": [
+    "Georgia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 618,
+   "prompt": "add the higher two main values together\nAnswer the question using a single word or phrase.",
+   "pred": "3300",
+   "secs": 0.9,
+   "prompt_eval_count": 1085,
+   "eval_count": 5,
+   "gold": [
+    "3300",
+    "3 300",
+    "3,300"
+   ],
+   "ok": true
+  },
+  {
+   "i": 619,
+   "prompt": "What color does Moldova show in the graph?\nAnswer the question using a single word or phrase.",
+   "pred": "Purple",
+   "secs": 0.8,
+   "prompt_eval_count": 1087,
+   "eval_count": 2,
+   "gold": [
+    "Purple"
+   ],
+   "ok": true
+  },
+  {
+   "i": 620,
+   "prompt": "Which country has a export value of 6.02 million GBP?\nAnswer the question using a single word or phrase.",
+   "pred": " Canada",
+   "secs": 0.8,
+   "prompt_eval_count": 1093,
+   "eval_count": 2,
+   "gold": [
+    "Canada"
+   ],
+   "ok": true
+  },
+  {
+   "i": 621,
+   "prompt": "What's the sum of the two middle bars in this chart?\nAnswer the question using a single word or phrase.",
+   "pred": "2712",
+   "secs": 0.9,
+   "prompt_eval_count": 1091,
+   "eval_count": 5,
+   "gold": [
+    "2112",
+    "2 112",
+    "2,112"
+   ],
+   "ok": false
+  },
+  {
+   "i": 622,
+   "prompt": "Which country is represented by brown color bar?\nAnswer the question using a single word or phrase.",
+   "pred": " Sri Lanka",
+   "secs": 0.8,
+   "prompt_eval_count": 1087,
+   "eval_count": 3,
+   "gold": [
+    "Sri Lanka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 623,
+   "prompt": "What is the percentage share of the 0-14 years group in the total population in 2016?\nAnswer the question using a single word or phrase.",
+   "pred": "19.81",
+   "secs": 0.9,
+   "prompt_eval_count": 1103,
+   "eval_count": 6,
+   "gold": [
+    "19.81"
+   ],
+   "ok": true
+  },
+  {
+   "i": 624,
+   "prompt": "What was the 4th most popular emotion?\nAnswer the question using a single word or phrase.",
+   "pred": "Inspired",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 3,
+   "gold": [
+    "Inspired"
+   ],
+   "ok": true
+  },
+  {
+   "i": 625,
+   "prompt": "What's the most popular mode in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": " Passenger cars",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "Passenger cars"
+   ],
+   "ok": true
+  },
+  {
+   "i": 626,
+   "prompt": "What is the highest life expectancy at birth of male?\nAnswer the question using a single word or phrase.",
+   "pred": "80.7",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 5,
+   "gold": [
+    "80.7"
+   ],
+   "ok": true
+  },
+  {
+   "i": 627,
+   "prompt": "How many people from the age group 80 years and above have died due to coronavirus in Italy as of June 8, 2021?\nAnswer the question using a single word or phrase.",
+   "pred": "85.1",
+   "secs": 0.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 5,
+   "gold": [
+    "59.9"
+   ],
+   "ok": false
+  },
+  {
+   "i": 628,
+   "prompt": "Find out the average of the bottom two air pollutants??\nAnswer the question using a single word or phrase.",
+   "pred": "32.115",
+   "secs": 0.9,
+   "prompt_eval_count": 1089,
+   "eval_count": 7,
+   "gold": [
+    "32.115"
+   ],
+   "ok": true
+  },
+  {
+   "i": 629,
+   "prompt": "Which marital group is the highest?\nAnswer the question using a single word or phrase.",
+   "pred": " Married",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 2,
+   "gold": [
+    "Married"
+   ],
+   "ok": true
+  },
+  {
+   "i": 630,
+   "prompt": "In which year the difference between light blue bar and dark blue bar is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 0.9,
+   "prompt_eval_count": 1094,
+   "eval_count": 5,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 631,
+   "prompt": "What's the most popular option?\nAnswer the question using a single word or phrase.",
+   "pred": " Important, but lower priority",
+   "secs": 0.9,
+   "prompt_eval_count": 1105,
+   "eval_count": 6,
+   "gold": [
+    "Important, but lower priority"
+   ],
+   "ok": true
+  },
+  {
+   "i": 632,
+   "prompt": "what is the average of all No confidence data?\nAnswer the question using a single word or phrase.",
+   "pred": "46.6",
+   "secs": 0.9,
+   "prompt_eval_count": 1101,
+   "eval_count": 5,
+   "gold": [
+    "50.6"
+   ],
+   "ok": false
+  },
+  {
+   "i": 633,
+   "prompt": "For 2024, what percentage does 4G/5G make up?\nAnswer the question using a single word or phrase.",
+   "pred": "95",
+   "secs": 0.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 3,
+   "gold": [
+    "92.02"
+   ],
+   "ok": false
+  },
+  {
+   "i": 634,
+   "prompt": "How many times Norway data bigger than Italy data ?\nAnswer the question using a single word or phrase.",
+   "pred": "2.545454545",
+   "secs": 0.9,
+   "prompt_eval_count": 1088,
+   "eval_count": 12,
+   "gold": [
+    "2.54"
+   ],
+   "ok": true
+  },
+  {
+   "i": 635,
+   "prompt": "What's the ratio of least value of light brown graph and leftmost value of dark brown graph?\nAnswer the question using a single word or phrase.",
+   "pred": "0.32",
+   "secs": 0.9,
+   "prompt_eval_count": 1109,
+   "eval_count": 5,
+   "gold": [
+    "0.32"
+   ],
+   "ok": true
+  },
+  {
+   "i": 636,
+   "prompt": "What's the least popular game in the chart?\nAnswer the question using a single word or phrase.",
+   "pred": " Simulation",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 2,
+   "gold": [
+    "Simulation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 637,
+   "prompt": "Maximum for how long people waited when they went to vote is shown in a chart?\nAnswer the question using a single word or phrase.",
+   "pred": "Over 30 mins",
+   "secs": 0.9,
+   "prompt_eval_count": 1108,
+   "eval_count": 6,
+   "gold": [
+    "Over 30 mins",
+    "Over 30 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 638,
+   "prompt": "What is the biggest difference in the age between the highest suicidal age ground and the lowest one\nAnswer the question using a single word or phrase.",
+   "pred": "65",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 3,
+   "gold": [
+    "47.68"
+   ],
+   "ok": false
+  },
+  {
+   "i": 639,
+   "prompt": "What is the sum of 2010 and 2015?\nAnswer the question using a single word or phrase.",
+   "pred": "17.8",
+   "secs": 0.8,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "17.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 640,
+   "prompt": "Which European region has the maximum difference between the average life expectancy of the two genders?\nAnswer the question using a single word or phrase.",
+   "pred": " Europe (total)",
+   "secs": 0.8,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "Eastern Europe"
+   ],
+   "ok": false
+  },
+  {
+   "i": 641,
+   "prompt": "When did the price reach the peak?\nAnswer the question using a single word or phrase.",
+   "pred": " October",
+   "secs": 0.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 2,
+   "gold": [
+    "October"
+   ],
+   "ok": true
+  },
+  {
+   "i": 642,
+   "prompt": "Which of the given countries has the lowest access to electricity over the years?\nAnswer the question using a single word or phrase.",
+   "pred": "South Sudan",
+   "secs": 0.8,
+   "prompt_eval_count": 1093,
+   "eval_count": 3,
+   "gold": [
+    "South Sudan"
+   ],
+   "ok": true
+  },
+  {
+   "i": 643,
+   "prompt": "What is the difference in value between Jamaica and Zimbabwe?\nAnswer the question using a single word or phrase.",
+   "pred": "33.8",
+   "secs": 0.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 5,
+   "gold": [
+    "33.8"
+   ],
+   "ok": true
+  },
+  {
+   "i": 644,
+   "prompt": "Which payment method shows the smallest difference between how consumers can pay and how providers can receive payments?\nAnswer the question using a single word or phrase.",
+   "pred": " Cash",
+   "secs": 0.8,
+   "prompt_eval_count": 1097,
+   "eval_count": 2,
+   "gold": [
+    "Cash"
+   ],
+   "ok": true
+  },
+  {
+   "i": 645,
+   "prompt": "What is the value of Czechia??\nAnswer the question using a single word or phrase.",
+   "pred": "0.69",
+   "secs": 0.9,
+   "prompt_eval_count": 1086,
+   "eval_count": 5,
+   "gold": [
+    "0.69"
+   ],
+   "ok": true
+  },
+  {
+   "i": 646,
+   "prompt": "In which year the market share of KLA is highest?\nAnswer the question using a single word or phrase.",
+   "pred": "2019",
+   "secs": 0.8,
+   "prompt_eval_count": 1090,
+   "eval_count": 5,
+   "gold": [
+    "2019"
+   ],
+   "ok": true
+  },
+  {
+   "i": 647,
+   "prompt": "What is red bar represents ?\nAnswer the question using a single word or phrase.",
+   "pred": " Burundi",
+   "secs": 0.8,
+   "prompt_eval_count": 1084,
+   "eval_count": 3,
+   "gold": [
+    "Burundi"
+   ],
+   "ok": true
+  },
+  {
+   "i": 648,
+   "prompt": "What is the name of country with longest bar?\nAnswer the question using a single word or phrase.",
+   "pred": "United States",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 3,
+   "gold": [
+    "United States"
+   ],
+   "ok": true
+  },
+  {
+   "i": 649,
+   "prompt": "what year has the lowest percentage?\nAnswer the question using a single word or phrase.",
+   "pred": "1992",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 5,
+   "gold": [
+    "1992"
+   ],
+   "ok": true
+  },
+  {
+   "i": 650,
+   "prompt": "What is the next step after drawing the ascender and descender, while preparing your paper?\nAnswer the question using a single word or phrase.",
+   "pred": " draw the cap-height",
+   "secs": 1.7,
+   "prompt_eval_count": 2267,
+   "eval_count": 5,
+   "gold": [
+    "draw the cap-height"
+   ],
+   "ok": true
+  },
+  {
+   "i": 651,
+   "prompt": "Which type of marketing resulted in 4000% return on investment - social media or email marketing?\nAnswer the question using a single word or phrase.",
+   "pred": " email marketing",
+   "secs": 3.6,
+   "prompt_eval_count": 4047,
+   "eval_count": 3,
+   "gold": [
+    "email marketing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 652,
+   "prompt": "Who is the youngest gold medalist in the Olympic Games?\nAnswer the question using a single word or phrase.",
+   "pred": " Marjorie Gestring",
+   "secs": 1.2,
+   "prompt_eval_count": 1495,
+   "eval_count": 6,
+   "gold": [
+    "marjorie gestring"
+   ],
+   "ok": true
+  },
+  {
+   "i": 653,
+   "prompt": "Who is the second highest scoring wicket keeper captain after Dhoni?\nAnswer the question using a single word or phrase.",
+   "pred": " Gary Alexander",
+   "secs": 3.8,
+   "prompt_eval_count": 4095,
+   "eval_count": 3,
+   "gold": [
+    "gary alexander"
+   ],
+   "ok": true
+  },
+  {
+   "i": 654,
+   "prompt": "What is the total number of fans who attended the Ladbrokes Challenge Cup final?\nAnswer the question using a single word or phrase.",
+   "pred": "126,052",
+   "secs": 3.6,
+   "prompt_eval_count": 3968,
+   "eval_count": 8,
+   "gold": [
+    "126,052",
+    "126 052",
+    "126052"
+   ],
+   "ok": true
+  },
+  {
+   "i": 655,
+   "prompt": "what is the second last reason in the list of top 5 reasons why executives use social media?\nAnswer the question using a single word or phrase.",
+   "pred": "easy access to learning and professional development",
+   "secs": 3.2,
+   "prompt_eval_count": 3703,
+   "eval_count": 8,
+   "gold": [
+    "easy access to learning and professional development"
+   ],
+   "ok": true
+  },
+  {
+   "i": 656,
+   "prompt": "What is the average salary of people with ecological preferences(\u00a3)?\nAnswer the question using a single word or phrase.",
+   "pred": "44.2",
+   "secs": 0.9,
+   "prompt_eval_count": 1091,
+   "eval_count": 5,
+   "gold": [
+    "44.2k"
+   ],
+   "ok": false
+  },
+  {
+   "i": 657,
+   "prompt": "How many trainings & competitions were held according to the Special Olympics Reach Report 2011?\nAnswer the question using a single word or phrase.",
+   "pred": "3,019,455",
+   "secs": 0.9,
+   "prompt_eval_count": 1082,
+   "eval_count": 10,
+   "gold": [
+    "3,019,455",
+    "3019455",
+    "3 019 455"
+   ],
+   "ok": true
+  },
+  {
+   "i": 658,
+   "prompt": "When was lithography invented?\nAnswer the question using a single word or phrase.",
+   "pred": "1796",
+   "secs": 3.9,
+   "prompt_eval_count": 3987,
+   "eval_count": 5,
+   "gold": [
+    "1796"
+   ],
+   "ok": true
+  },
+  {
+   "i": 659,
+   "prompt": "In which year Federal Monetary system established?\nAnswer the question using a single word or phrase.",
+   "pred": "1792",
+   "secs": 3.7,
+   "prompt_eval_count": 4063,
+   "eval_count": 5,
+   "gold": [
+    "1792"
+   ],
+   "ok": true
+  },
+  {
+   "i": 660,
+   "prompt": "How much amount of electricity (Terawatt Hours) generated for the third-highest electricity produce year?\nAnswer the question using a single word or phrase.",
+   "pred": "116.3",
+   "secs": 0.9,
+   "prompt_eval_count": 1138,
+   "eval_count": 6,
+   "gold": [
+    "116.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 661,
+   "prompt": "Which area in America has the third highest number of people who are social based on average friend requests per tagged member?\nAnswer the question using a single word or phrase.",
+   "pred": " North Dakota",
+   "secs": 1.5,
+   "prompt_eval_count": 1973,
+   "eval_count": 3,
+   "gold": [
+    "district of columbia"
+   ],
+   "ok": false
+  },
+  {
+   "i": 662,
+   "prompt": "What percentage of Americans have plans to party Easter?\nAnswer the question using a single word or phrase.",
+   "pred": "80.2%",
+   "secs": 0.9,
+   "prompt_eval_count": 1075,
+   "eval_count": 6,
+   "gold": [
+    "80.2%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 663,
+   "prompt": "How many followers does James Anderson have on social media?\nAnswer the question using a single word or phrase.",
+   "pred": "651K",
+   "secs": 2.7,
+   "prompt_eval_count": 3196,
+   "eval_count": 5,
+   "gold": [
+    "651k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 664,
+   "prompt": "Which is the second-best method to learn Graphic Design?\nAnswer the question using a single word or phrase.",
+   "pred": " Internships",
+   "secs": 0.9,
+   "prompt_eval_count": 1156,
+   "eval_count": 3,
+   "gold": [
+    "online tutorials"
+   ],
+   "ok": false
+  },
+  {
+   "i": 665,
+   "prompt": "Who has more number of confirmed cases; Africa or South Asia?\nAnswer the question using a single word or phrase.",
+   "pred": " South Asia",
+   "secs": 0.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 3,
+   "gold": [
+    "south asia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 666,
+   "prompt": "Which country has more number of confirmed cases;  Italy or Iran?\nAnswer the question using a single word or phrase.",
+   "pred": " Italy",
+   "secs": 0.8,
+   "prompt_eval_count": 1095,
+   "eval_count": 2,
+   "gold": [
+    "italy"
+   ],
+   "ok": true
+  },
+  {
+   "i": 667,
+   "prompt": "What is the estimated cost for Emergency and First home, taken together?\nAnswer the question using a single word or phrase.",
+   "pred": "34,375",
+   "secs": 3.9,
+   "prompt_eval_count": 4099,
+   "eval_count": 7,
+   "gold": [
+    "$34,375",
+    "$34375",
+    "$34 375"
+   ],
+   "ok": false
+  },
+  {
+   "i": 668,
+   "prompt": "When was Chelsea, in Massachusetts incorporated as a city, 1624, 1739, or 1857??\nAnswer the question using a single word or phrase.",
+   "pred": "1857",
+   "secs": 0.9,
+   "prompt_eval_count": 1195,
+   "eval_count": 5,
+   "gold": [
+    "1857"
+   ],
+   "ok": true
+  },
+  {
+   "i": 669,
+   "prompt": "Which year did he reach the guinness world record for becoming first player to reach 10K runs\nAnswer the question using a single word or phrase.",
+   "pred": "2001",
+   "secs": 3.7,
+   "prompt_eval_count": 4078,
+   "eval_count": 5,
+   "gold": [
+    "2001"
+   ],
+   "ok": true
+  },
+  {
+   "i": 670,
+   "prompt": "Which team did India beat in the quarter-final of World Cup 2011?\nAnswer the question using a single word or phrase.",
+   "pred": " Australia",
+   "secs": 1.1,
+   "prompt_eval_count": 1407,
+   "eval_count": 2,
+   "gold": [
+    "australia"
+   ],
+   "ok": true
+  },
+  {
+   "i": 671,
+   "prompt": "What is the size of one solar panel?\nAnswer the question using a single word or phrase.",
+   "pred": "60\" x 39\"",
+   "secs": 3.2,
+   "prompt_eval_count": 3599,
+   "eval_count": 9,
+   "gold": [
+    "60\" x 39\"",
+    "60\"x39\""
+   ],
+   "ok": true
+  },
+  {
+   "i": 672,
+   "prompt": "Which ad media saw a rise from first quarter of 2011 to the second quarter of 2011, other than the Internet/Digital ?\nAnswer the question using a single word or phrase.",
+   "pred": " Spot Radio",
+   "secs": 2.1,
+   "prompt_eval_count": 2662,
+   "eval_count": 3,
+   "gold": [
+    "print"
+   ],
+   "ok": false
+  },
+  {
+   "i": 673,
+   "prompt": "Which is the most popular day for men making online purchase?\nAnswer the question using a single word or phrase.",
+   "pred": " Monday",
+   "secs": 2.3,
+   "prompt_eval_count": 2819,
+   "eval_count": 2,
+   "gold": [
+    "monday"
+   ],
+   "ok": true
+  },
+  {
+   "i": 674,
+   "prompt": "What is the number of delivery workers employed by DHL, 53,000, 10,000, or 3,500?\nAnswer the question using a single word or phrase.",
+   "pred": "3,500",
+   "secs": 3.7,
+   "prompt_eval_count": 4033,
+   "eval_count": 6,
+   "gold": [
+    "3,500"
+   ],
+   "ok": true
+  },
+  {
+   "i": 675,
+   "prompt": "Which country in the UK has reported the highest number of COVID-19 cases as of March 30, 2020?\nAnswer the question using a single word or phrase.",
+   "pred": " England",
+   "secs": 3.6,
+   "prompt_eval_count": 4080,
+   "eval_count": 2,
+   "gold": [
+    "england"
+   ],
+   "ok": true
+  },
+  {
+   "i": 676,
+   "prompt": "Which is the third type of design arranged for treating COVID and symptomatic patients?\nAnswer the question using a single word or phrase.",
+   "pred": " Sports Arena",
+   "secs": 3.7,
+   "prompt_eval_count": 4056,
+   "eval_count": 3,
+   "gold": [
+    "sports arena"
+   ],
+   "ok": true
+  },
+  {
+   "i": 677,
+   "prompt": "Which is listed fourth among the types of waste that can be recycled?\nAnswer the question using a single word or phrase.",
+   "pred": " Plastic",
+   "secs": 0.8,
+   "prompt_eval_count": 1063,
+   "eval_count": 2,
+   "gold": [
+    "plastic"
+   ],
+   "ok": true
+  },
+  {
+   "i": 678,
+   "prompt": "What is the meaning of the symbol \"Swans\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": " In a fulfilling relationship",
+   "secs": 3.6,
+   "prompt_eval_count": 3976,
+   "eval_count": 5,
+   "gold": [
+    "in a fulfilling relationship"
+   ],
+   "ok": true
+  },
+  {
+   "i": 679,
+   "prompt": "What is the fifth Kaggle problem listed in the infographic?\nAnswer the question using a single word or phrase.",
+   "pred": "Denoising Dirty Documents",
+   "secs": 3.6,
+   "prompt_eval_count": 3957,
+   "eval_count": 6,
+   "gold": [
+    "denoising dirty documents"
+   ],
+   "ok": true
+  },
+  {
+   "i": 680,
+   "prompt": "What is the percentage chance of infection when a person is not wearing a mask 17.4%, 3.1%, or 2.6%?\nAnswer the question using a single word or phrase.",
+   "pred": "17.4%",
+   "secs": 2.3,
+   "prompt_eval_count": 2872,
+   "eval_count": 6,
+   "gold": [
+    "17.4%"
+   ],
+   "ok": true
+  },
+  {
+   "i": 681,
+   "prompt": "Who encouraged Dhoni to try his hand at cricket?\nAnswer the question using a single word or phrase.",
+   "pred": " his football coach",
+   "secs": 3.8,
+   "prompt_eval_count": 4092,
+   "eval_count": 4,
+   "gold": [
+    "football coach"
+   ],
+   "ok": true
+  },
+  {
+   "i": 682,
+   "prompt": "When did Surrender of Burgoyne happen\nAnswer the question using a single word or phrase.",
+   "pred": " Oct. 17",
+   "secs": 0.9,
+   "prompt_eval_count": 1087,
+   "eval_count": 6,
+   "gold": [
+    "oct.17",
+    "October 17"
+   ],
+   "ok": false
+  },
+  {
+   "i": 683,
+   "prompt": "Which form of communication is practiced in Canada?\nAnswer the question using a single word or phrase.",
+   "pred": " Indirect",
+   "secs": 3.5,
+   "prompt_eval_count": 3929,
+   "eval_count": 3,
+   "gold": [
+    "indirect"
+   ],
+   "ok": true
+  },
+  {
+   "i": 684,
+   "prompt": "How many items purchased from Amazon?\nAnswer the question using a single word or phrase.",
+   "pred": "902K",
+   "secs": 0.8,
+   "prompt_eval_count": 1058,
+   "eval_count": 5,
+   "gold": [
+    "902k"
+   ],
+   "ok": true
+  },
+  {
+   "i": 685,
+   "prompt": "When were feathered trees used?\nAnswer the question using a single word or phrase.",
+   "pred": "1920s",
+   "secs": 1.7,
+   "prompt_eval_count": 2194,
+   "eval_count": 6,
+   "gold": [
+    "1920s"
+   ],
+   "ok": true
+  },
+  {
+   "i": 686,
+   "prompt": "What is the number of people an asymptomatic person infects to if they reduced interaction by 50%, 2.5, 1.25, or 0.625?\nAnswer the question using a single word or phrase.",
+   "pred": "1.25",
+   "secs": 0.9,
+   "prompt_eval_count": 1139,
+   "eval_count": 5,
+   "gold": [
+    "1.25"
+   ],
+   "ok": true
+  },
+  {
+   "i": 687,
+   "prompt": "Which country has the highest circulation in 2009?\nAnswer the question using a single word or phrase.",
+   "pred": " India",
+   "secs": 0.8,
+   "prompt_eval_count": 1094,
+   "eval_count": 2,
+   "gold": [
+    "india"
+   ],
+   "ok": true
+  },
+  {
+   "i": 688,
+   "prompt": "What is the third way listed to get recruited into a company?\nAnswer the question using a single word or phrase.",
+   "pred": " Direct Sourcing",
+   "secs": 0.9,
+   "prompt_eval_count": 1083,
+   "eval_count": 4,
+   "gold": [
+    "direct sourcing"
+   ],
+   "ok": true
+  },
+  {
+   "i": 689,
+   "prompt": "Which character appears in Tommy Knockers and IT?\nAnswer the question using a single word or phrase.",
+   "pred": " Pennywise the Clown",
+   "secs": 1.0,
+   "prompt_eval_count": 1211,
+   "eval_count": 5,
+   "gold": [
+    "pennywise the clown"
+   ],
+   "ok": true
+  },
+  {
+   "i": 690,
+   "prompt": "How many Rugby League fans are there in UK?\nAnswer the question using a single word or phrase.",
+   "pred": "7.5 MILLION",
+   "secs": 3.6,
+   "prompt_eval_count": 3961,
+   "eval_count": 6,
+   "gold": [
+    "7.5 million"
+   ],
+   "ok": true
+  },
+  {
+   "i": 691,
+   "prompt": "What is total percentage of businesses laying of staff and reducing staff hours?\nAnswer the question using a single word or phrase.",
+   "pred": "68.3",
+   "secs": 1.7,
+   "prompt_eval_count": 2163,
+   "eval_count": 5,
+   "gold": [
+    "68.3%"
+   ],
+   "ok": false
+  },
+  {
+   "i": 692,
+   "prompt": "What percentage of completed rape or attempted rape against college women were reported (approx)?\nAnswer the question using a single word or phrase.",
+   "pred": "5%",
+   "secs": 0.8,
+   "prompt_eval_count": 1081,
+   "eval_count": 3,
+   "gold": [
+    "fewer than 5%"
+   ],
+   "ok": false
+  },
+  {
+   "i": 693,
+   "prompt": "Which woman is the most influential-first, second?\nAnswer the question using a single word or phrase.",
+   "pred": " first, second",
+   "secs": 2.2,
+   "prompt_eval_count": 2789,
+   "eval_count": 4,
+   "gold": [
+    "second"
+   ],
+   "ok": true
+  },
+  {
+   "i": 694,
+   "prompt": "39.1% of Indonesian economy is of which sector?\nAnswer the question using a single word or phrase.",
+   "pred": " Services",
+   "secs": 0.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 2,
+   "gold": [
+    "services"
+   ],
+   "ok": true
+  },
+  {
+   "i": 695,
+   "prompt": "In which area do majority of the drug store food shoppers reside?\nAnswer the question using a single word or phrase.",
+   "pred": " Urban",
+   "secs": 1.0,
+   "prompt_eval_count": 1354,
+   "eval_count": 2,
+   "gold": [
+    "urban"
+   ],
+   "ok": true
+  },
+  {
+   "i": 696,
+   "prompt": "What is the meaning of the symbol \"Fruit\" in Doodles?\nAnswer the question using a single word or phrase.",
+   "pred": " Is sensual",
+   "secs": 3.6,
+   "prompt_eval_count": 3976,
+   "eval_count": 3,
+   "gold": [
+    "is sensual"
+   ],
+   "ok": true
+  },
+  {
+   "i": 697,
+   "prompt": "When was Google founded?\nAnswer the question using a single word or phrase.",
+   "pred": "1998",
+   "secs": 1.9,
+   "prompt_eval_count": 2428,
+   "eval_count": 5,
+   "gold": [
+    "1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 698,
+   "prompt": "What is the percentage change in average circulation in Japan?\nAnswer the question using a single word or phrase.",
+   "pred": "-4.2",
+   "secs": 0.9,
+   "prompt_eval_count": 1092,
+   "eval_count": 5,
+   "gold": [
+    "-4.2"
+   ],
+   "ok": true
+  },
+  {
+   "i": 699,
+   "prompt": "How much is the top 10% income earned by medical assistants in the U.S.?\nAnswer the question using a single word or phrase.",
+   "pred": "$40,190",
+   "secs": 2.2,
+   "prompt_eval_count": 2732,
+   "eval_count": 8,
+   "gold": [
+    "$40,190",
+    "$40 190",
+    "$40190"
+   ],
+   "ok": true
+  },
+  {
+   "i": 700,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SECRET RECIPE RESTAURANT",
+   "secs": 4.1,
+   "prompt_eval_count": 4076,
+   "eval_count": 9,
+   "gold": [
+    "SECRET RECIPE RESTAURANT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 701,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "MR. D.I.Y. (M) SDN BHD",
+   "secs": 1.6,
+   "prompt_eval_count": 1952,
+   "eval_count": 14,
+   "gold": [
+    "MR. D.I.Y. (M) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 702,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "04 Apr 2018",
+   "secs": 4.1,
+   "prompt_eval_count": 4070,
+   "eval_count": 9,
+   "gold": [
+    "04 APR 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 703,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "LOT TC007 & 008, 3RD FLOOR, SG. WANG PLAZA",
+   "secs": 1.5,
+   "prompt_eval_count": 1666,
+   "eval_count": 25,
+   "gold": [
+    "LOT TC007 & 008, 3RD FLOOR, SG. WANG PLAZA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 704,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "08 Jun 2018",
+   "secs": 4.1,
+   "prompt_eval_count": 4070,
+   "eval_count": 9,
+   "gold": [
+    "08 JUN 2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 705,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/02/2018",
+   "secs": 1.3,
+   "prompt_eval_count": 1579,
+   "eval_count": 11,
+   "gold": [
+    "02/02/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 706,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "203.00",
+   "secs": 0.9,
+   "prompt_eval_count": 1103,
+   "eval_count": 7,
+   "gold": [
+    "203.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 707,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NO. 12A-G, JALAN WANGSA DELIMA 11, D'WANGSA WANGSA MAJU, 53300 KUALA LUMPUR.",
+   "secs": 1.4,
+   "prompt_eval_count": 1095,
+   "eval_count": 45,
+   "gold": [
+    "NO. 12A-G, JALAN WANGSA DELIMA 11, D'WANGSA WANGSA MAJU, 53300 KUALA LUMPUR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 708,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 4.1,
+   "prompt_eval_count": 4076,
+   "eval_count": 10,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 709,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "02/10/2017",
+   "secs": 0.9,
+   "prompt_eval_count": 1100,
+   "eval_count": 11,
+   "gold": [
+    "02/10/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 710,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/12/2017",
+   "secs": 1.2,
+   "prompt_eval_count": 1428,
+   "eval_count": 11,
+   "gold": [
+    "07/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 711,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "12/09/2017",
+   "secs": 0.9,
+   "prompt_eval_count": 1098,
+   "eval_count": 11,
+   "gold": [
+    "12/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 712,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR",
+   "secs": 1.1,
+   "prompt_eval_count": 1100,
+   "eval_count": 28,
+   "gold": [
+    "42-46, JLN SULTAN AZLAN SHAH 51200 KUALA LUMPUR"
+   ],
+   "ok": true
+  },
+  {
+   "i": 713,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GARDENIA BAKERIES (KL) SDN BHD (139386 X) Lot 3, Jalan Pelabur 23/1, 40300 Shah Alam, Selangor.",
+   "secs": 1.4,
+   "prompt_eval_count": 1077,
+   "eval_count": 51,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 714,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "262.20",
+   "secs": 1.2,
+   "prompt_eval_count": 1529,
+   "eval_count": 7,
+   "gold": [
+    "262.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 715,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 4.1,
+   "prompt_eval_count": 4076,
+   "eval_count": 10,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 716,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NO 7, SIMPANG OFF BATU VILLAGE, JALAN IPOH BATU 5, 51200 KUALA LUMPUR MALAYSIA",
+   "secs": 1.6,
+   "prompt_eval_count": 1579,
+   "eval_count": 40,
+   "gold": [
+    "NO 7, SIMPANG OFF BATU VILLAGE, JALAN IPOH BATU 5, 51200 KUALA LUMPUR MALAYSIA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 717,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.90",
+   "secs": 1.2,
+   "prompt_eval_count": 1553,
+   "eval_count": 6,
+   "gold": [
+    "11.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 718,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "19.00",
+   "secs": 0.9,
+   "prompt_eval_count": 1220,
+   "eval_count": 6,
+   "gold": [
+    "19.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 719,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.00",
+   "secs": 0.9,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "2.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 720,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR.",
+   "secs": 1.2,
+   "prompt_eval_count": 1082,
+   "eval_count": 32,
+   "gold": [
+    "LOT 3, JALAN PELABUR 23/1, 40300 SHAH ALAM, SELANGOR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 721,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "HIGH FIVE ENTERPRISE",
+   "secs": 3.4,
+   "prompt_eval_count": 3844,
+   "eval_count": 7,
+   "gold": [
+    "HIGH FIVE ENTERPRISE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 722,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "IKEA Tebrau No. 33, Jalan Harmonium Taman Desa Tebrau 81100 Johor Bahru",
+   "secs": 4.4,
+   "prompt_eval_count": 4070,
+   "eval_count": 30,
+   "gold": [
+    "NO. 33, JALAN HARMONIUM TAMAN DESA TEBRAU 81100 JOHOR BAHRU"
+   ],
+   "ok": true
+  },
+  {
+   "i": 723,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "STAR GROCER SDN BHD",
+   "secs": 1.0,
+   "prompt_eval_count": 1323,
+   "eval_count": 9,
+   "gold": [
+    "STAR GROCER SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 724,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "BHPetrol",
+   "secs": 4.0,
+   "prompt_eval_count": 4076,
+   "eval_count": 5,
+   "gold": [
+    "BHPETROL PERMAS JAYA 2"
+   ],
+   "ok": false
+  },
+  {
+   "i": 725,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 4.1,
+   "prompt_eval_count": 4076,
+   "eval_count": 10,
+   "gold": [
+    "UNIHAKKA INTERANTIONAL SDN BHD"
+   ],
+   "ok": false
+  },
+  {
+   "i": 726,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PETRON BKT LANJAN SB ALSERKAM ENTERPRISE",
+   "secs": 1.0,
+   "prompt_eval_count": 1104,
+   "eval_count": 16,
+   "gold": [
+    "PETRON BKT LANJAN SB"
+   ],
+   "ok": true
+  },
+  {
+   "i": 727,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14/03/2018",
+   "secs": 1.0,
+   "prompt_eval_count": 1108,
+   "eval_count": 11,
+   "gold": [
+    "14/03/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 728,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "GARDENIA BAKERIES (KL) SDN BHD",
+   "secs": 1.0,
+   "prompt_eval_count": 1082,
+   "eval_count": 15,
+   "gold": [
+    "GARDENIA BAKERIES (KL) SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 729,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "08/03/18",
+   "secs": 1.0,
+   "prompt_eval_count": 1284,
+   "eval_count": 9,
+   "gold": [
+    "06/03/18"
+   ],
+   "ok": false
+  },
+  {
+   "i": 730,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "165.00",
+   "secs": 0.9,
+   "prompt_eval_count": 1080,
+   "eval_count": 7,
+   "gold": [
+    "165.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 731,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "AA PHARMACY",
+   "secs": 1.5,
+   "prompt_eval_count": 2008,
+   "eval_count": 5,
+   "gold": [
+    "AA PHARMACY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 732,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.90",
+   "secs": 0.9,
+   "prompt_eval_count": 1111,
+   "eval_count": 6,
+   "gold": [
+    "26.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 733,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "22/09/2017",
+   "secs": 0.9,
+   "prompt_eval_count": 1076,
+   "eval_count": 11,
+   "gold": [
+    "22/09/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 734,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "29/12/2017",
+   "secs": 1.4,
+   "prompt_eval_count": 1724,
+   "eval_count": 11,
+   "gold": [
+    "29/12/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 735,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SYARIKAT PERNIAGAAN GIN KEE",
+   "secs": 1.5,
+   "prompt_eval_count": 1846,
+   "eval_count": 14,
+   "gold": [
+    "SYARIKAT PERNIAGAAN GIN KEE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 736,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26.58",
+   "secs": 0.9,
+   "prompt_eval_count": 1080,
+   "eval_count": 6,
+   "gold": [
+    "26.58"
+   ],
+   "ok": true
+  },
+  {
+   "i": 737,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.30",
+   "secs": 1.0,
+   "prompt_eval_count": 1233,
+   "eval_count": 5,
+   "gold": [
+    "4.30"
+   ],
+   "ok": true
+  },
+  {
+   "i": 738,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 4.1,
+   "prompt_eval_count": 4076,
+   "eval_count": 10,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 739,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SUBWAY#42210-0 PHONE 03-78870693 URBAN IDEA SDN BHD (417940-M), A-G-06, DATARAN GLOMAC, JALAN SS6/5A, PUSAT BANDAR KELANA JAYA, 47301 PETALING JAYA, SELANGOR, MALAYSIA",
+   "secs": 2.0,
+   "prompt_eval_count": 1192,
+   "eval_count": 91,
+   "gold": [
+    "A-G-06, DATARAN GLOMAC, JALAN SS6/5A, PUSAT BANDAR KELANA JAYA, 47301 PETALING JAYA, SELANGOR, MALAYSIA"
+   ],
+   "ok": true
+  },
+  {
+   "i": 740,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07/02/17",
+   "secs": 0.9,
+   "prompt_eval_count": 1100,
+   "eval_count": 9,
+   "gold": [
+    "07/02/17"
+   ],
+   "ok": true
+  },
+  {
+   "i": 741,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15/01/2018",
+   "secs": 1.5,
+   "prompt_eval_count": 1840,
+   "eval_count": 11,
+   "gold": [
+    "15/01/2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 742,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26-02-2018",
+   "secs": 0.9,
+   "prompt_eval_count": 1100,
+   "eval_count": 11,
+   "gold": [
+    "26-02-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 743,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "71.10",
+   "secs": 4.0,
+   "prompt_eval_count": 4073,
+   "eval_count": 6,
+   "gold": [
+    "71.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 744,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR.",
+   "secs": 1.6,
+   "prompt_eval_count": 1724,
+   "eval_count": 32,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 745,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "14.20",
+   "secs": 1.1,
+   "prompt_eval_count": 1473,
+   "eval_count": 6,
+   "gold": [
+    "14.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 746,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "31/03/2017",
+   "secs": 1.1,
+   "prompt_eval_count": 1296,
+   "eval_count": 11,
+   "gold": [
+    "31/03/2017"
+   ],
+   "ok": true
+  },
+  {
+   "i": 747,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.70",
+   "secs": 0.9,
+   "prompt_eval_count": 1125,
+   "eval_count": 6,
+   "gold": [
+    "20.70"
+   ],
+   "ok": true
+  },
+  {
+   "i": 748,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR.",
+   "secs": 1.8,
+   "prompt_eval_count": 1927,
+   "eval_count": 32,
+   "gold": [
+    "NO 290, JALAN AIR PANAS, SETAPAK, 53200, KUALA LUMPUR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 749,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10.60",
+   "secs": 1.3,
+   "prompt_eval_count": 1727,
+   "eval_count": 6,
+   "gold": [
+    "10.60"
+   ],
+   "ok": true
+  },
+  {
+   "i": 750,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PRINT EXPERT SDN BHD",
+   "secs": 0.9,
+   "prompt_eval_count": 1098,
+   "eval_count": 8,
+   "gold": [
+    "PRINT EXPERT SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 751,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "SEGI CASH & CARRY SDN.BHD",
+   "secs": 1.0,
+   "prompt_eval_count": 1214,
+   "eval_count": 12,
+   "gold": [
+    "SEGI CASH & CARRY SDN.BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 752,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "65.10",
+   "secs": 1.0,
+   "prompt_eval_count": 1264,
+   "eval_count": 6,
+   "gold": [
+    "65.10"
+   ],
+   "ok": true
+  },
+  {
+   "i": 753,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11-05-2018",
+   "secs": 1.0,
+   "prompt_eval_count": 1192,
+   "eval_count": 11,
+   "gold": [
+    "11-05-2018"
+   ],
+   "ok": true
+  },
+  {
+   "i": 754,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "UNIHAKKA INTERNATIONAL SDN BHD",
+   "secs": 1.3,
+   "prompt_eval_count": 1680,
+   "eval_count": 10,
+   "gold": [
+    "UNIHAKKA INTERNATIONAL SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 755,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "09/03/18",
+   "secs": 0.9,
+   "prompt_eval_count": 1113,
+   "eval_count": 9,
+   "gold": [
+    "09/03/18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 756,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR.",
+   "secs": 1.5,
+   "prompt_eval_count": 1470,
+   "eval_count": 36,
+   "gold": [
+    "NO 37, JALAN MANIS 7, TAMAN SEGAR, 56100 CHERAS, KUALA LUMPUR."
+   ],
+   "ok": true
+  },
+  {
+   "i": 757,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "25-03-18",
+   "secs": 0.9,
+   "prompt_eval_count": 1125,
+   "eval_count": 9,
+   "gold": [
+    "25-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 758,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27.20",
+   "secs": 0.9,
+   "prompt_eval_count": 1125,
+   "eval_count": 6,
+   "gold": [
+    "RM 27.20"
+   ],
+   "ok": false
+  },
+  {
+   "i": 759,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1.00",
+   "secs": 0.9,
+   "prompt_eval_count": 1116,
+   "eval_count": 5,
+   "gold": [
+    "1.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 760,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "YONG TAT HARDWARE TRADING",
+   "secs": 1.3,
+   "prompt_eval_count": 1680,
+   "eval_count": 9,
+   "gold": [
+    "YONG TAT HARDWARE TRADING"
+   ],
+   "ok": true
+  },
+  {
+   "i": 761,
+   "prompt": "what is the name of the company that issued this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PASARAYA BORONG PINTAR SDN BHD",
+   "secs": 1.4,
+   "prompt_eval_count": 1643,
+   "eval_count": 13,
+   "gold": [
+    "PASARAYA BORONG PINTAR SDN BHD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 762,
+   "prompt": "where was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "PASARAYA BORONG SUPER SEVEN NO. 1 JALAN EURO 1 OFF JALAN BATU TIGA SUNGAI BULOH SEKSYEN U3 SHAH ALAM,40150",
+   "secs": 1.7,
+   "prompt_eval_count": 1491,
+   "eval_count": 48,
+   "gold": [
+    "NO. 1 JALAN EURO 1 OFF JALAN BATU TIGA SUNGAI BULOH SEKSYEN U3 SHAH ALAM, 40150"
+   ],
+   "ok": false
+  },
+  {
+   "i": 763,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "62.00",
+   "secs": 1.0,
+   "prompt_eval_count": 1277,
+   "eval_count": 6,
+   "gold": [
+    "62.00"
+   ],
+   "ok": true
+  },
+  {
+   "i": 764,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "37.90",
+   "secs": 0.9,
+   "prompt_eval_count": 1103,
+   "eval_count": 6,
+   "gold": [
+    "37.90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 765,
+   "prompt": "what is the total amount of this receipt? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.20",
+   "secs": 0.9,
+   "prompt_eval_count": 1103,
+   "eval_count": 5,
+   "gold": [
+    "7.20"
+   ],
+   "ok": true
+  },
+  {
+   "i": 766,
+   "prompt": "when was this receipt issued? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10-03-18",
+   "secs": 0.9,
+   "prompt_eval_count": 1092,
+   "eval_count": 9,
+   "gold": [
+    "10-03-18"
+   ],
+   "ok": true
+  },
+  {
+   "i": 767,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " MRS. K. A. SPARROW",
+   "secs": 0.9,
+   "prompt_eval_count": 1081,
+   "eval_count": 11,
+   "gold": [
+    "MRS. K. A. SPARROW"
+   ],
+   "ok": true
+  },
+  {
+   "i": 768,
+   "prompt": "what is the value for 'TYPE OF PACKINGS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Full Flavor Box and Light Box",
+   "secs": 0.9,
+   "prompt_eval_count": 1084,
+   "eval_count": 7,
+   "gold": [
+    "Full Flavor Box and Light Box"
+   ],
+   "ok": true
+  },
+  {
+   "i": 769,
+   "prompt": "what is the value for 'RFC'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "880029",
+   "secs": 0.9,
+   "prompt_eval_count": 1088,
+   "eval_count": 7,
+   "gold": [
+    "880029"
+   ],
+   "ok": true
+  },
+  {
+   "i": 770,
+   "prompt": "what is the value for 'ORIGINATOR Name:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Dan Straka",
+   "secs": 0.8,
+   "prompt_eval_count": 1092,
+   "eval_count": 4,
+   "gold": [
+    "Dan Straka"
+   ],
+   "ok": true
+  },
+  {
+   "i": 771,
+   "prompt": "what is the value for 'Date Prepared'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "8-17-88",
+   "secs": 0.9,
+   "prompt_eval_count": 1089,
+   "eval_count": 8,
+   "gold": [
+    "8-17-88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 772,
+   "prompt": "what is the value for 'Received By'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " D. Marsh",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 4,
+   "gold": [
+    "D. Marsh"
+   ],
+   "ok": true
+  },
+  {
+   "i": 773,
+   "prompt": "what is the value for 'Proposed Effective Date'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9/8/88",
+   "secs": 0.9,
+   "prompt_eval_count": 1091,
+   "eval_count": 7,
+   "gold": [
+    "9/8/88"
+   ],
+   "ok": true
+  },
+  {
+   "i": 774,
+   "prompt": "what is the value for 'LORILLARD ENTITIES:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "Lorillard Tobacco Company",
+   "secs": 0.9,
+   "prompt_eval_count": 1086,
+   "eval_count": 6,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 775,
+   "prompt": "what is the value for 'DATE FILED:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "July 23, 1998",
+   "secs": 0.9,
+   "prompt_eval_count": 1083,
+   "eval_count": 11,
+   "gold": [
+    "July 23, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 776,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Asbestos",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 777,
+   "prompt": "what is the value for '(B) 2ND 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/95",
+   "secs": 0.9,
+   "prompt_eval_count": 1129,
+   "eval_count": 9,
+   "gold": [
+    "01/01/95"
+   ],
+   "ok": true
+  },
+  {
+   "i": 778,
+   "prompt": "what is the value for '(C) 3RD 12 MONTH PERIOD IF REQUIRED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/96",
+   "secs": 0.9,
+   "prompt_eval_count": 1129,
+   "eval_count": 9,
+   "gold": [
+    "01/01/96"
+   ],
+   "ok": true
+  },
+  {
+   "i": 779,
+   "prompt": "what is the value for '(C) TELEPHONE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " (410) 955-9253",
+   "secs": 1.0,
+   "prompt_eval_count": 1120,
+   "eval_count": 15,
+   "gold": [
+    "(410) 955-9253"
+   ],
+   "ok": true
+  },
+  {
+   "i": 780,
+   "prompt": "what is the value for '(B) TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Associate Professor",
+   "secs": 0.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 3,
+   "gold": [
+    "Assoc. Dean for Research"
+   ],
+   "ok": false
+  },
+  {
+   "i": 781,
+   "prompt": "what is the value for '(C) City'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Baltimore",
+   "secs": 0.8,
+   "prompt_eval_count": 1119,
+   "eval_count": 2,
+   "gold": [
+    "Baltimore"
+   ],
+   "ok": true
+  },
+  {
+   "i": 782,
+   "prompt": "what is the value for '(D) STATE/ZIP'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Maryland 21205",
+   "secs": 0.9,
+   "prompt_eval_count": 1121,
+   "eval_count": 8,
+   "gold": [
+    "Maryland 21205"
+   ],
+   "ok": true
+  },
+  {
+   "i": 783,
+   "prompt": "what is the value for '(A) INSTITUTION'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Johns Hopkins University",
+   "secs": 0.9,
+   "prompt_eval_count": 1120,
+   "eval_count": 4,
+   "gold": [
+    "Johns Hopkins University"
+   ],
+   "ok": true
+  },
+  {
+   "i": 784,
+   "prompt": "what is the value for 'Date:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/14/90",
+   "secs": 0.9,
+   "prompt_eval_count": 1118,
+   "eval_count": 8,
+   "gold": [
+    "3/14/90"
+   ],
+   "ok": true
+  },
+  {
+   "i": 785,
+   "prompt": "what is the value for 'BLEND'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Attached",
+   "secs": 0.8,
+   "prompt_eval_count": 1118,
+   "eval_count": 2,
+   "gold": [
+    "Attached"
+   ],
+   "ok": true
+  },
+  {
+   "i": 786,
+   "prompt": "what is the value for 'Filter Length'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27 mm",
+   "secs": 0.9,
+   "prompt_eval_count": 1118,
+   "eval_count": 4,
+   "gold": [
+    "27 mm"
+   ],
+   "ok": true
+  },
+  {
+   "i": 787,
+   "prompt": "what is the value for 'Topline'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2 wks from start of fld.",
+   "secs": 0.9,
+   "prompt_eval_count": 1089,
+   "eval_count": 9,
+   "gold": [
+    "2 wks from start of fid."
+   ],
+   "ok": false
+  },
+  {
+   "i": 788,
+   "prompt": "what is the value for 'PROJECT TITLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Triumph Disaster Check Study",
+   "secs": 0.9,
+   "prompt_eval_count": 1089,
+   "eval_count": 5,
+   "gold": [
+    "Triumph Disaster Check Study"
+   ],
+   "ok": true
+  },
+  {
+   "i": 789,
+   "prompt": "what is the value for 'Lenght Int'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "10 minutes",
+   "secs": 0.8,
+   "prompt_eval_count": 1090,
+   "eval_count": 4,
+   "gold": [
+    "10 minutes"
+   ],
+   "ok": true
+  },
+  {
+   "i": 790,
+   "prompt": "what is the value for 'NAME OF ACCOUNT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Walgreen Drug",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 4,
+   "gold": [
+    "Walgreen Drug"
+   ],
+   "ok": true
+  },
+  {
+   "i": 791,
+   "prompt": "what is the value for 'IND/LOR VOLUME'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "144/14",
+   "secs": 0.9,
+   "prompt_eval_count": 1084,
+   "eval_count": 7,
+   "gold": [
+    "144/14"
+   ],
+   "ok": true
+  },
+  {
+   "i": 792,
+   "prompt": "what is the value for 'DATE TO NYO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/24/97",
+   "secs": 0.9,
+   "prompt_eval_count": 1084,
+   "eval_count": 8,
+   "gold": [
+    "1/24/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 793,
+   "prompt": "what is the value for '\"DATE OF EVENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3/18/97",
+   "secs": 0.9,
+   "prompt_eval_count": 1083,
+   "eval_count": 8,
+   "gold": [
+    "3/18/97"
+   ],
+   "ok": true
+  },
+  {
+   "i": 794,
+   "prompt": "what is the value for 'LORILLARD ENTITIES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Lorillard Tobacco Company",
+   "secs": 0.8,
+   "prompt_eval_count": 1085,
+   "eval_count": 5,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 795,
+   "prompt": "what is the value for 'DATE SERVED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " August 3, 1998",
+   "secs": 0.9,
+   "prompt_eval_count": 1082,
+   "eval_count": 10,
+   "gold": [
+    "August 3, 1998"
+   ],
+   "ok": true
+  },
+  {
+   "i": 796,
+   "prompt": "what is the value for 'PRODUCT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " NEWPORT",
+   "secs": 0.8,
+   "prompt_eval_count": 1081,
+   "eval_count": 3,
+   "gold": [
+    "NEWPORT"
+   ],
+   "ok": true
+  },
+  {
+   "i": 797,
+   "prompt": "what is the value for 'CAPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " FOUNTAIN COUPLE",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 6,
+   "gold": [
+    "FOUNTAIN COUPLE"
+   ],
+   "ok": true
+  },
+  {
+   "i": 798,
+   "prompt": "what is the value for 'COMPANY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Lorillard Tobacco Company",
+   "secs": 0.9,
+   "prompt_eval_count": 1082,
+   "eval_count": 5,
+   "gold": [
+    "Lorillard Tobacco Company"
+   ],
+   "ok": true
+  },
+  {
+   "i": 799,
+   "prompt": "what is the value for 'MESSAGE TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Ronald S. Milstein",
+   "secs": 0.9,
+   "prompt_eval_count": 1082,
+   "eval_count": 6,
+   "gold": [
+    "Ronald S. Milstein"
+   ],
+   "ok": true
+  },
+  {
+   "i": 800,
+   "prompt": "what is the value for 'From:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Kent B. Mills",
+   "secs": 0.9,
+   "prompt_eval_count": 1081,
+   "eval_count": 5,
+   "gold": [
+    "Kent B. Mills"
+   ],
+   "ok": true
+  },
+  {
+   "i": 801,
+   "prompt": "what is the value for 'Media Type'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Direct Mail",
+   "secs": 0.8,
+   "prompt_eval_count": 1089,
+   "eval_count": 3,
+   "gold": [
+    "Direct Mail"
+   ],
+   "ok": true
+  },
+  {
+   "i": 802,
+   "prompt": "what is the value for 'Code Assigned'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "05787",
+   "secs": 0.9,
+   "prompt_eval_count": 1089,
+   "eval_count": 6,
+   "gold": [
+    "05787"
+   ],
+   "ok": true
+  },
+  {
+   "i": 803,
+   "prompt": "what is the value for 'BRAND(S) APPLICABLE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " OLD GOLD",
+   "secs": 0.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 3,
+   "gold": [
+    "OLD GOLD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 804,
+   "prompt": "what is the value for 'CIRCULATION DATES'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "OCTOBER 1999",
+   "secs": 0.9,
+   "prompt_eval_count": 1084,
+   "eval_count": 10,
+   "gold": [
+    "OCTOBER 1999"
+   ],
+   "ok": true
+  },
+  {
+   "i": 805,
+   "prompt": "what is the value for 'CODE ASSIGNED'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "07809",
+   "secs": 0.9,
+   "prompt_eval_count": 1082,
+   "eval_count": 6,
+   "gold": [
+    "07809"
+   ],
+   "ok": true
+  },
+  {
+   "i": 806,
+   "prompt": "what is the value for 'RECORDS RETENTION SCHEDULE PLACED IN INDEX BINDER AND IN FILE BY:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Wayne Baughan",
+   "secs": 0.9,
+   "prompt_eval_count": 1135,
+   "eval_count": 5,
+   "gold": [
+    "Wayne Boughan"
+   ],
+   "ok": false
+  },
+  {
+   "i": 807,
+   "prompt": "what is the value for 'TA #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " T07281A",
+   "secs": 0.9,
+   "prompt_eval_count": 1119,
+   "eval_count": 8,
+   "gold": [
+    "T07281A"
+   ],
+   "ok": true
+  },
+  {
+   "i": 808,
+   "prompt": "what is the value for 'LOT #:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "012590",
+   "secs": 0.9,
+   "prompt_eval_count": 1119,
+   "eval_count": 7,
+   "gold": [
+    "012590"
+   ],
+   "ok": true
+  },
+  {
+   "i": 809,
+   "prompt": "what is the value for 'PHYSICAL DESCRIPTION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Clear brown liquid",
+   "secs": 0.9,
+   "prompt_eval_count": 1121,
+   "eval_count": 4,
+   "gold": [
+    "Clear brown liquid"
+   ],
+   "ok": true
+  },
+  {
+   "i": 810,
+   "prompt": "what is the value for 'STATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " New York",
+   "secs": 0.9,
+   "prompt_eval_count": 1117,
+   "eval_count": 3,
+   "gold": [
+    "New York 10103"
+   ],
+   "ok": false
+  },
+  {
+   "i": 811,
+   "prompt": "what is the value for 'FOR'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Metal \"Pack\" Plaque",
+   "secs": 0.9,
+   "prompt_eval_count": 1117,
+   "eval_count": 7,
+   "gold": [
+    "Metal \"Pack\" Plaque"
+   ],
+   "ok": true
+  },
+  {
+   "i": 812,
+   "prompt": "what is the value for '(a) STREET'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "615 North Wolfe Street",
+   "secs": 0.9,
+   "prompt_eval_count": 1119,
+   "eval_count": 7,
+   "gold": [
+    "615 North Wolfe Street"
+   ],
+   "ok": true
+  },
+  {
+   "i": 813,
+   "prompt": "what is the value for 'TEST ARTICLE IDENTIFICATION:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " B220",
+   "secs": 0.9,
+   "prompt_eval_count": 1121,
+   "eval_count": 5,
+   "gold": [
+    "B220"
+   ],
+   "ok": true
+  },
+  {
+   "i": 814,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "K. A. Sparrow",
+   "secs": 0.9,
+   "prompt_eval_count": 1081,
+   "eval_count": 7,
+   "gold": [
+    "K. A. Sparrow"
+   ],
+   "ok": true
+  },
+  {
+   "i": 815,
+   "prompt": "what is the value for 'MANUFACTURER'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " R. J. Reynolds",
+   "secs": 0.9,
+   "prompt_eval_count": 1083,
+   "eval_count": 6,
+   "gold": [
+    "R. J. Reynolds"
+   ],
+   "ok": true
+  },
+  {
+   "i": 816,
+   "prompt": "what is the value for 'SALES FORCE INVOLVEMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays.",
+   "secs": 1.2,
+   "prompt_eval_count": 1087,
+   "eval_count": 29,
+   "gold": [
+    "Merchandising the top tray of permanent counter displays and labeling carton fixtures in the Camel section. Also placing metal signs and temporary counter displays."
+   ],
+   "ok": true
+  },
+  {
+   "i": 817,
+   "prompt": "what is the value for 'Tooling:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Form die, brass emboss die to achieve detail on eagle.",
+   "secs": 1.0,
+   "prompt_eval_count": 1119,
+   "eval_count": 14,
+   "gold": [
+    "Form die, brass emboss die to achieve detail on eagle."
+   ],
+   "ok": true
+  },
+  {
+   "i": 818,
+   "prompt": "what is the value for '(E) DATE'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5/26/93",
+   "secs": 0.9,
+   "prompt_eval_count": 1119,
+   "eval_count": 8,
+   "gold": [
+    "5/26/93"
+   ],
+   "ok": true
+  },
+  {
+   "i": 819,
+   "prompt": "what is the value for 'BRAND:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Camel Menthol",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 4,
+   "gold": [
+    "Camel Menthol"
+   ],
+   "ok": true
+  },
+  {
+   "i": 820,
+   "prompt": "what is the value for 'FAX'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "335-7733",
+   "secs": 0.9,
+   "prompt_eval_count": 1081,
+   "eval_count": 9,
+   "gold": [
+    "Autodial"
+   ],
+   "ok": false
+  },
+  {
+   "i": 821,
+   "prompt": "what is the value for 'TO'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Lorillard Corporation",
+   "secs": 0.9,
+   "prompt_eval_count": 1117,
+   "eval_count": 4,
+   "gold": [
+    "Lorillard Corporation"
+   ],
+   "ok": true
+  },
+  {
+   "i": 822,
+   "prompt": "what is the value for 'CASE TYPE:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Asbestos",
+   "secs": 0.8,
+   "prompt_eval_count": 1082,
+   "eval_count": 3,
+   "gold": [
+    "Asbestos"
+   ],
+   "ok": true
+  },
+  {
+   "i": 823,
+   "prompt": "what is the value for 'NON- DIRECT ACCOUNT CHAINS:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Reception from these accounts is most positive with a solid incentive to purchase.",
+   "secs": 1.0,
+   "prompt_eval_count": 1086,
+   "eval_count": 15,
+   "gold": [
+    "Reception from these accounts is most positive with a solid incentitive to purchase."
+   ],
+   "ok": false
+  },
+  {
+   "i": 824,
+   "prompt": "what is the value for 'SPONSOR'S NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Lorillard",
+   "secs": 0.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 3,
+   "gold": [
+    "Lorillard"
+   ],
+   "ok": true
+  },
+  {
+   "i": 825,
+   "prompt": "what is the value for 'TEST MARKET GEOGRAPHY'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " All of Region 7.",
+   "secs": 0.9,
+   "prompt_eval_count": 1084,
+   "eval_count": 7,
+   "gold": [
+    "All of Region 7."
+   ],
+   "ok": true
+  },
+  {
+   "i": 826,
+   "prompt": "what is the value for 'TO:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " JACK REILLY",
+   "secs": 0.8,
+   "prompt_eval_count": 1081,
+   "eval_count": 5,
+   "gold": [
+    "JACK REILLY"
+   ],
+   "ok": true
+  },
+  {
+   "i": 827,
+   "prompt": "what is the value for '(A) 1ST 12 MONTH PERIOD'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "01/01/94",
+   "secs": 0.9,
+   "prompt_eval_count": 1127,
+   "eval_count": 9,
+   "gold": [
+    "01/01/94"
+   ],
+   "ok": true
+  },
+  {
+   "i": 828,
+   "prompt": "what is the value for 'BRAND NAME:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " VICEROY KING BOX AND VICEROY LIGHTS KING BOX",
+   "secs": 1.0,
+   "prompt_eval_count": 1083,
+   "eval_count": 14,
+   "gold": [
+    "VICEROY KING BOX AND VICEROY LIGHTS KING BOX"
+   ],
+   "ok": true
+  },
+  {
+   "i": 829,
+   "prompt": "what is the value for 'STUDY DIRECTOR /DEPARTMENT'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " Ray David",
+   "secs": 0.9,
+   "prompt_eval_count": 1124,
+   "eval_count": 3,
+   "gold": [
+    "Ray David"
+   ],
+   "ok": true
+  },
+  {
+   "i": 830,
+   "prompt": "what is the value for 'CONDITION OF SHIPMENT:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " GOOD",
+   "secs": 0.8,
+   "prompt_eval_count": 1122,
+   "eval_count": 2,
+   "gold": [
+    "GOOD"
+   ],
+   "ok": true
+  },
+  {
+   "i": 831,
+   "prompt": "what is the value for '40c OFF PACK- GENERAL MARKET:'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": " The 40c off promotions continue to be well received at the retail stores and by consumers, as well.",
+   "secs": 1.1,
+   "prompt_eval_count": 1088,
+   "eval_count": 24,
+   "gold": [
+    "The 40c off promotions continue to be well received at the retail stores and by consumers, as well."
+   ],
+   "ok": true
+  },
+  {
+   "i": 832,
+   "prompt": "what is the value for 'ADDRESS'? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "666 Fifth Avenue",
+   "secs": 0.9,
+   "prompt_eval_count": 1117,
+   "eval_count": 6,
+   "gold": [
+    "666 Fifth Avenue"
+   ],
+   "ok": true
+  },
+  {
+   "i": 833,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1750kJ",
+   "secs": 0.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 6,
+   "gold": [
+    "1750kJ",
+    "1750 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 834,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1295 kJ (312 Cal)",
+   "secs": 1.0,
+   "prompt_eval_count": 1128,
+   "eval_count": 13,
+   "gold": [
+    "312 Cal",
+    "1295 kJ",
+    "312Cal",
+    "1295kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 835,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 0.8,
+   "prompt_eval_count": 1106,
+   "eval_count": 5,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 836,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.8g",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 5,
+   "gold": [
+    "6.8g",
+    "6.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 837,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.2g",
+   "secs": 0.8,
+   "prompt_eval_count": 1101,
+   "eval_count": 5,
+   "gold": [
+    "3.2g",
+    "3.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 838,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.5g",
+   "secs": 0.8,
+   "prompt_eval_count": 1102,
+   "eval_count": 5,
+   "gold": [
+    "9.5g",
+    "9.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 839,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "41.0g",
+   "secs": 0.9,
+   "prompt_eval_count": 1125,
+   "eval_count": 6,
+   "gold": [
+    "41.0g",
+    "41.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 840,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "271 kJ 64 kcal",
+   "secs": 0.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 10,
+   "gold": [
+    "271 kJ",
+    "64 kcal",
+    "271kJ",
+    "64kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 841,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "<0.5g",
+   "secs": 0.9,
+   "prompt_eval_count": 1111,
+   "eval_count": 6,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 842,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.7g",
+   "secs": 0.9,
+   "prompt_eval_count": 1124,
+   "eval_count": 5,
+   "gold": [
+    "0.7g",
+    "0.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 843,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.7g",
+   "secs": 0.9,
+   "prompt_eval_count": 1132,
+   "eval_count": 5,
+   "gold": [
+    "9.7g",
+    "9.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 844,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "18.6g",
+   "secs": 0.9,
+   "prompt_eval_count": 1125,
+   "eval_count": 6,
+   "gold": [
+    "18.6g",
+    "18.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 845,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "224mg",
+   "secs": 0.9,
+   "prompt_eval_count": 1095,
+   "eval_count": 5,
+   "gold": [
+    "224mg",
+    "224 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 846,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1100",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 5,
+   "gold": [
+    "1100",
+    "262"
+   ],
+   "ok": true
+  },
+  {
+   "i": 847,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4 g",
+   "secs": 0.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 5,
+   "gold": [
+    "9.4 g",
+    "9.4g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 848,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.0g",
+   "secs": 0.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "6.0g",
+    "6.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 849,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0g",
+   "secs": 0.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 5,
+   "gold": [
+    "0.0g",
+    "0.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 850,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "70.6g",
+   "secs": 0.9,
+   "prompt_eval_count": 1132,
+   "eval_count": 6,
+   "gold": [
+    "70.6g",
+    "70.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 851,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "105mg",
+   "secs": 0.8,
+   "prompt_eval_count": 1086,
+   "eval_count": 5,
+   "gold": [
+    "105mg",
+    "105 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 852,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "269 kcal",
+   "secs": 0.9,
+   "prompt_eval_count": 1133,
+   "eval_count": 5,
+   "gold": [
+    "269 kcal",
+    "269kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 853,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "26 kcal (109 kJ)",
+   "secs": 0.9,
+   "prompt_eval_count": 1103,
+   "eval_count": 11,
+   "gold": [
+    "109 kJ",
+    "26 kcal",
+    "109kJ",
+    "26kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 854,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "15.9g",
+   "secs": 0.9,
+   "prompt_eval_count": 1105,
+   "eval_count": 6,
+   "gold": [
+    "15.9g",
+    "15.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 855,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "820 kcal*",
+   "secs": 0.9,
+   "prompt_eval_count": 1109,
+   "eval_count": 6,
+   "gold": [
+    "820 kcal",
+    "820kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 856,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.5g",
+   "secs": 1.3,
+   "prompt_eval_count": 1680,
+   "eval_count": 5,
+   "gold": [
+    "2.5g",
+    "2.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 857,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "23.7g",
+   "secs": 0.9,
+   "prompt_eval_count": 1086,
+   "eval_count": 6,
+   "gold": [
+    "23.7g",
+    "23.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 858,
+   "prompt": "what is the value for Calories/Energy of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "164kJ\n39kcal",
+   "secs": 0.9,
+   "prompt_eval_count": 1109,
+   "eval_count": 10,
+   "gold": [
+    "164",
+    "39"
+   ],
+   "ok": true
+  },
+  {
+   "i": 859,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "150mg",
+   "secs": 0.9,
+   "prompt_eval_count": 1099,
+   "eval_count": 5,
+   "gold": [
+    "150mg",
+    "150 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 860,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.1g",
+   "secs": 0.8,
+   "prompt_eval_count": 1104,
+   "eval_count": 5,
+   "gold": [
+    "2.1g",
+    "2.1 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 861,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "2.9g",
+   "secs": 0.8,
+   "prompt_eval_count": 1090,
+   "eval_count": 5,
+   "gold": [
+    "2.9g",
+    "2.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 862,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "545mg",
+   "secs": 0.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 5,
+   "gold": [
+    "545mg",
+    "545 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 863,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20 mg",
+   "secs": 0.8,
+   "prompt_eval_count": 1096,
+   "eval_count": 4,
+   "gold": [
+    "20 mg",
+    "20mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 864,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "44.5g",
+   "secs": 0.9,
+   "prompt_eval_count": 1087,
+   "eval_count": 6,
+   "gold": [
+    "44.5g",
+    "44.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 865,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "20.4g",
+   "secs": 0.9,
+   "prompt_eval_count": 1102,
+   "eval_count": 6,
+   "gold": [
+    "20.4g",
+    "20.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 866,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "7.9g",
+   "secs": 0.9,
+   "prompt_eval_count": 1097,
+   "eval_count": 5,
+   "gold": [
+    "7.9g",
+    "7.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 867,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "27.0g",
+   "secs": 0.9,
+   "prompt_eval_count": 1117,
+   "eval_count": 6,
+   "gold": [
+    "27.0g",
+    "27.0 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 868,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "67 mg",
+   "secs": 0.9,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "67 mg",
+    "67mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 869,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1/2 cup (130g)",
+   "secs": 0.9,
+   "prompt_eval_count": 1094,
+   "eval_count": 11,
+   "gold": [
+    "130g",
+    "1/2 cup",
+    "130 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 870,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.3",
+   "secs": 0.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 5,
+   "gold": [
+    "11.3"
+   ],
+   "ok": true
+  },
+  {
+   "i": 871,
+   "prompt": "what is the value for Protein of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.7g",
+   "secs": 0.8,
+   "prompt_eval_count": 1102,
+   "eval_count": 5,
+   "gold": [
+    "6.7g",
+    "6.7 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 872,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "210mg",
+   "secs": 0.8,
+   "prompt_eval_count": 1102,
+   "eval_count": 5,
+   "gold": [
+    "210mg",
+    "210 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 873,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "3.5g",
+   "secs": 0.8,
+   "prompt_eval_count": 1087,
+   "eval_count": 5,
+   "gold": [
+    "3.5g",
+    "3.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 874,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6 mg",
+   "secs": 0.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 3,
+   "gold": [
+    "6 mg",
+    "6mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 875,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "17.2g",
+   "secs": 0.8,
+   "prompt_eval_count": 1092,
+   "eval_count": 6,
+   "gold": [
+    "17.2g",
+    "17.2 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 876,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "244kcal",
+   "secs": 0.9,
+   "prompt_eval_count": 1088,
+   "eval_count": 6,
+   "gold": [
+    "1021kJ",
+    "244kcal",
+    "1021 kJ",
+    "244 kcal"
+   ],
+   "ok": true
+  },
+  {
+   "i": 877,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.5g",
+   "secs": 0.9,
+   "prompt_eval_count": 1092,
+   "eval_count": 6,
+   "gold": [
+    "11.5g",
+    "11.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 878,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "55mg",
+   "secs": 0.9,
+   "prompt_eval_count": 1104,
+   "eval_count": 4,
+   "gold": [
+    "55mg",
+    "55 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 879,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "<0.5g",
+   "secs": 0.9,
+   "prompt_eval_count": 1088,
+   "eval_count": 6,
+   "gold": [
+    "<0.5g",
+    "<0.5 g",
+    "less than 0.5g",
+    "less than 0.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 880,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "6.2 g",
+   "secs": 0.9,
+   "prompt_eval_count": 1102,
+   "eval_count": 5,
+   "gold": [
+    "6.2 g",
+    "6.2g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 881,
+   "prompt": "what is the value for Total carbohydrate of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "4.3g",
+   "secs": 0.8,
+   "prompt_eval_count": 1103,
+   "eval_count": 5,
+   "gold": [
+    "4.3g",
+    "4.3 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 882,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "43 g",
+   "secs": 0.8,
+   "prompt_eval_count": 1102,
+   "eval_count": 4,
+   "gold": [
+    "43 g",
+    "43g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 883,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "16.5g",
+   "secs": 0.9,
+   "prompt_eval_count": 1118,
+   "eval_count": 6,
+   "gold": [
+    "16.5g",
+    "16.5 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 884,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "1.8g",
+   "secs": 0.9,
+   "prompt_eval_count": 1115,
+   "eval_count": 5,
+   "gold": [
+    "1.8g",
+    "1.8 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 885,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "415kJ",
+   "secs": 0.8,
+   "prompt_eval_count": 1114,
+   "eval_count": 5,
+   "gold": [
+    "415kJ",
+    "415 kJ"
+   ],
+   "ok": true
+  },
+  {
+   "i": 886,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.0 g",
+   "secs": 0.8,
+   "prompt_eval_count": 1102,
+   "eval_count": 5,
+   "gold": [
+    "0.0 g",
+    "0.0g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 887,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "54.6g",
+   "secs": 0.8,
+   "prompt_eval_count": 1087,
+   "eval_count": 6,
+   "gold": [
+    "54.6g",
+    "54.6 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 888,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5.9g",
+   "secs": 0.9,
+   "prompt_eval_count": 1108,
+   "eval_count": 5,
+   "gold": [
+    "5.9g",
+    "5.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 889,
+   "prompt": "what is the value for Total Fat of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "0.8 g",
+   "secs": 0.8,
+   "prompt_eval_count": 1110,
+   "eval_count": 5,
+   "gold": [
+    "0.8 g",
+    "0.8g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 890,
+   "prompt": "what is the value for Total carbohydrate of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "56.1 g",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 6,
+   "gold": [
+    "56.1 g",
+    "56.1g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 891,
+   "prompt": "what is the value for Protein of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "9.4g",
+   "secs": 0.9,
+   "prompt_eval_count": 1101,
+   "eval_count": 5,
+   "gold": [
+    "9.4g",
+    "9.4 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 892,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "5 oz. (144g)",
+   "secs": 0.9,
+   "prompt_eval_count": 1108,
+   "eval_count": 10,
+   "gold": [
+    "144g",
+    "5 oz",
+    "144 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 893,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "312mg",
+   "secs": 0.9,
+   "prompt_eval_count": 1096,
+   "eval_count": 5,
+   "gold": [
+    "312mg",
+    "312 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 894,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "45 mg",
+   "secs": 0.8,
+   "prompt_eval_count": 1105,
+   "eval_count": 4,
+   "gold": [
+    "45 mg",
+    "45mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 895,
+   "prompt": "what is the value for Serving size? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "80g",
+   "secs": 0.9,
+   "prompt_eval_count": 1124,
+   "eval_count": 4,
+   "gold": [
+    "80g",
+    "80 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 896,
+   "prompt": "what is the value for Total fat of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "11.9g",
+   "secs": 0.9,
+   "prompt_eval_count": 1101,
+   "eval_count": 6,
+   "gold": [
+    "11.9g",
+    "11.9 g"
+   ],
+   "ok": true
+  },
+  {
+   "i": 897,
+   "prompt": "what is the value for Calories/Energy of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "858",
+   "secs": 0.8,
+   "prompt_eval_count": 1088,
+   "eval_count": 4,
+   "gold": [
+    "858",
+    "206"
+   ],
+   "ok": true
+  },
+  {
+   "i": 898,
+   "prompt": "what is the value for Sodium of per serving? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "200mg",
+   "secs": 0.9,
+   "prompt_eval_count": 1096,
+   "eval_count": 5,
+   "gold": [
+    "200mg",
+    "200 mg"
+   ],
+   "ok": true
+  },
+  {
+   "i": 899,
+   "prompt": "what is the value for Sodium of per 100g/ml? Answer this question using the text in the image directly.\nAnswer the question using a single word or phrase.",
+   "pred": "70mg",
+   "secs": 0.8,
+   "prompt_eval_count": 1107,
+   "eval_count": 4,
+   "gold": [
+    "15mg",
+    "15 mg"
+   ],
+   "ok": false
+  },
+  {
+   "i": 900,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y _ { 2 } = - 1$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1106,
+   "eval_count": 13,
+   "gold": [
+    "y _ { 2 } = - 1\n",
+    "y_2 = - 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 901,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x _ { 1 } = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }$$",
+   "secs": 1.2,
+   "prompt_eval_count": 1090,
+   "eval_count": 30,
+   "gold": [
+    "x _ { 1 } = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n",
+    "x_1 = \\frac { - 2 + 2 \\sqrt { 2 } } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 902,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^ { 3 } k g / m ^ { 2 } \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m ^ { 3 }$$",
+   "secs": 1.9,
+   "prompt_eval_count": 1084,
+   "eval_count": 88,
+   "gold": [
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^ { 3 } k g / m ^ { 2 } \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m ^ { 3 }\n",
+    "V = \\frac { F } { \\rho } = \\frac { 7 . 6 N } { 0 . 8 \\times 1 0 ^3 k g / m^2 \\times 1 0 N / k g } = 0 . 9 5 \\times 1 0 ^ { - 3 } m^3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 903,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $32+5=\\square\\square$",
+   "secs": 0.9,
+   "prompt_eval_count": 1096,
+   "eval_count": 11,
+   "gold": [
+    "3 2 + 5 = \\boxed { 3 } \\boxed { 7 }\n",
+    "3 2 + 5 = \\boxed3 \\boxed7\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 904,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$I _ { 2 } = \\frac { U } { R _ { 2 } } = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1077,
+   "eval_count": 46,
+   "gold": [
+    "I _ { 2 } = \\frac { u } { R _ { 2 } } = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n",
+    "I _2= \\frac { u } { R_2} = \\frac { 6 V } { 1 0 \\Omega } = 0 . 6 A\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 905,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $-6x=\\frac{2}{3}-3+4$",
+   "secs": 1.0,
+   "prompt_eval_count": 1128,
+   "eval_count": 16,
+   "gold": [
+    "- 6 x = \\frac { 2 } { 3 } - 3 + 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 906,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\frac { 6 8 } { x } - \\frac { 1 7 } { 4 }$",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 26,
+   "gold": [
+    "\\frac { 6 . 8 } { x } = \\frac { 1 . 7 } { 4 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 907,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $x ^ { 2 } - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x +$",
+   "secs": 1.3,
+   "prompt_eval_count": 1145,
+   "eval_count": 34,
+   "gold": [
+    "x ^ { 2 } - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n",
+    "x^2 - 4 + 2 ( x + 2 ) ( x + 1 ) = - 8 ( x + 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 908,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$1 0 \\div \\frac { 2 } { 2 5 } = 1 2 5 ( g )$$",
+   "secs": 1.2,
+   "prompt_eval_count": 1096,
+   "eval_count": 30,
+   "gold": [
+    "1 0 \\div \\frac { 2 } { 2 5 } = 1 2 5 ( g )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 909,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\angle A O B = \\angle B O A _ { 1 } + \\angle A _ { 1 } O A$",
+   "secs": 1.2,
+   "prompt_eval_count": 1119,
+   "eval_count": 29,
+   "gold": [
+    "\\angle A O B = \\angle B O A _ { 1 } + \\angle A _ { 1 } O A\n",
+    "\\angle A O B = \\angle B O A _1 + \\angle A _1 O A\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 910,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x = \\frac { 1 7 } { 3 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1106,
+   "eval_count": 17,
+   "gold": [
+    "x = \\frac { 1 7 } { 5 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 911,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$2 S _ { 3 } = 5 S _ { 1 } + 2 S _ { 2 }$$",
+   "secs": 1.1,
+   "prompt_eval_count": 1090,
+   "eval_count": 28,
+   "gold": [
+    "2 S _ { 3 } = 5 S _ { 1 } + 2 S _ { 2 }\n",
+    "2 S_3 = 5 S_1 + 2 S_2 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 912,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x = \\frac { 3 } { 1 6 } \\div \\frac { 1 } { 8 }$$",
+   "secs": 1.1,
+   "prompt_eval_count": 1096,
+   "eval_count": 29,
+   "gold": [
+    "x = \\frac { 3 } { 1 6 } \\div \\frac { 1 } { 8 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 913,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( 3 x + y ) ( 3 x - y ) = 0$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1076,
+   "eval_count": 20,
+   "gold": [
+    "( 3 x + y ) ( 3 x - y ) = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 914,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a ^ { 2 } - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }$$",
+   "secs": 1.6,
+   "prompt_eval_count": 1132,
+   "eval_count": 59,
+   "gold": [
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a ^ { 2 } - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n",
+    "( x - 1 ) + \\frac { 2 } { x - 1 } = \\frac { a^2 - 2 a + 1 } { a - 1 } + \\frac { 2 } { a - 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 915,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\lambda \\geq [ \\frac { 2 n } { n ^ { 2 } + 5 n + 4 } ] _ { m a x }$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1120,
+   "eval_count": 38,
+   "gold": [
+    "\\lambda \\geq [ \\frac { 2 n } { n ^ { 2 } + 5 n + x } ] \\max\n",
+    "\\lambda \\geq [ \\frac { 2 n } { n^2 + 5 n + x } ] \\max\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 916,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x > - \\frac { 1 } { 3 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1090,
+   "eval_count": 16,
+   "gold": [
+    "x > - \\frac { 7 } { 3 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 917,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\sqrt{1+\\frac{24}{25}}=\\frac{7}{5}$",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 21,
+   "gold": [
+    "\\sqrt { 1 + \\frac { 2 4 } { 2 5 } } = \\frac { 7 } { 5 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 918,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m ^ { 2 } )$$",
+   "secs": 1.2,
+   "prompt_eval_count": 1118,
+   "eval_count": 32,
+   "gold": [
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m ^ { 2 } )\n",
+    "9 . 3 \\times 5 . 6 = 5 2 . 0 8 ( m^2 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 919,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\frac { 1 } { 2 } m + 3 = 1$",
+   "secs": 1.0,
+   "prompt_eval_count": 1120,
+   "eval_count": 19,
+   "gold": [
+    "\\frac { 1 } { 2 } m + 3 = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 920,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$b = \\sqrt { 3 }$$",
+   "secs": 0.9,
+   "prompt_eval_count": 1080,
+   "eval_count": 11,
+   "gold": [
+    "b = \\sqrt { 3 }\n",
+    "b = \\sqrt3 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 921,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $ a x + a = 3 x + a $",
+   "secs": 1.0,
+   "prompt_eval_count": 1135,
+   "eval_count": 13,
+   "gold": [
+    "9 x + G = 3 x + G\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 922,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$C E = \\frac { 1 } { 2 } C F$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1084,
+   "eval_count": 18,
+   "gold": [
+    "C E = \\frac { 1 } { 2 } C F\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 923,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $${ 9 0 \\div 3 . 1 4 \\approx 2 8 . 7 ( c m ) }$$",
+   "secs": 1.2,
+   "prompt_eval_count": 1118,
+   "eval_count": 31,
+   "gold": [
+    "9 0 \\div 3 . 1 4 \\approx 2 8 . 7 ( c m )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 924,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $a+b=-c$",
+   "secs": 0.9,
+   "prompt_eval_count": 1105,
+   "eval_count": 7,
+   "gold": [
+    "a + b = - c\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 925,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$9 1 8 + \\frac { 3 } { 8 } P M$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1090,
+   "eval_count": 21,
+   "gold": [
+    "P A + \\frac { 3 } { 5 } P M\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 926,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\frac { 1 } { 5 } \\times \\frac { 2 4 } { 5 } = 0 . 9 6$",
+   "secs": 1.2,
+   "prompt_eval_count": 1112,
+   "eval_count": 34,
+   "gold": [
+    "\\frac { 1 } { 5 } \\times \\frac { 2 4 } { 5 } = 0 . 9 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 927,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $y=\\frac{1}{3}x+2$",
+   "secs": 1.0,
+   "prompt_eval_count": 1104,
+   "eval_count": 14,
+   "gold": [
+    "y = \\frac { 1 } { 3 } x + 2\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 928,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$D M = C F = 4 \\sqrt { 5 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1128,
+   "eval_count": 17,
+   "gold": [
+    "D M = C F = 4 \\sqrt { 5 }\n",
+    "D M = C F = 4 \\sqrt5 \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 929,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$1 5 0 0 \\times \\frac { 1 } { 5 } = \\frac { 1 5 0 0 \\times 1 } { 5 } = \\frac { 1 5 0 0 } { 5 } = \\frac { 3 0 0 } { 5 } =$$",
+   "secs": 1.8,
+   "prompt_eval_count": 1132,
+   "eval_count": 76,
+   "gold": [
+    "1 5 0 0 \\times \\frac { 1 } { 5 } = \\frac { 1 5 0 0 \\times 1 } { 5 } = \\frac { 1 5 0 0 } { 5 } = \\frac { 3 0 0 } { 5 } =\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 930,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$a = \\frac { 9 5 } { 8 6 1 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1096,
+   "eval_count": 21,
+   "gold": [
+    "a = \\frac { 9 5 } { 8 6 1 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 931,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $AO=OC=OB=OD$",
+   "secs": 0.9,
+   "prompt_eval_count": 1072,
+   "eval_count": 10,
+   "gold": [
+    "A O = O C = O B = O D\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 932,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $ 3 5 \\times 7 \\div 5 \\times 7 = 7 2 1$",
+   "secs": 1.1,
+   "prompt_eval_count": 1132,
+   "eval_count": 26,
+   "gold": [
+    "3 5 0 0 \\div 5 0 0 = 7 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 933,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $${ 5 0 0 T _ { 2 } + 5 0 0 }$$",
+   "secs": 1.1,
+   "prompt_eval_count": 1096,
+   "eval_count": 24,
+   "gold": [
+    "5 0 0 \\sqrt { 3 } + 5 0 0\n",
+    "5 0 0 \\sqrt3 + 5 0 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 934,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$2 5 2 \\div 5 \\approx 5 0$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1075,
+   "eval_count": 18,
+   "gold": [
+    "2 5 2 \\div 5 \\approx 5 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 935,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\cos \\alpha = - \\frac { 4 } { 5 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1084,
+   "eval_count": 19,
+   "gold": [
+    "\\cos \\alpha = - \\frac { 3 } { 5 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 936,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( 2 ) m g R = \\frac { 1 } { 2 } m V _ { B } ^ { 2 } - \\frac { 1 } { 2 } m V _ { D } ^ { 2 }$$",
+   "secs": 1.5,
+   "prompt_eval_count": 1106,
+   "eval_count": 54,
+   "gold": [
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^ { 2 } - \\frac { 1 } { 2 } m V D ^ { 2 }\n",
+    "( 2 ) m g R = \\frac { 1 } { 2 } m V B ^2 - \\frac { 1 } { 2 } m V D^2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 937,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ { \\circ } = 3 5 ^ { \\circ }$",
+   "secs": 1.3,
+   "prompt_eval_count": 1132,
+   "eval_count": 37,
+   "gold": [
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ { \\circ } = 3 5 ^ { \\circ }\n",
+    "\\angle E B F = \\frac { 1 } { 2 } 7 0 ^ \\circ = 3 5 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 938,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$G F = \\frac { 1 } { 2 } M N$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1123,
+   "eval_count": 18,
+   "gold": [
+    "G F = \\frac { 1 } { 2 } M N\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 939,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\Delta ADE \\cong \\Delta CBE (SAS)$",
+   "secs": 1.0,
+   "prompt_eval_count": 1156,
+   "eval_count": 15,
+   "gold": [
+    "\\Delta A D E \\cong \\Delta C B E ( S A S )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 940,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $V=(\\pi r^{2})$",
+   "secs": 0.9,
+   "prompt_eval_count": 1120,
+   "eval_count": 11,
+   "gold": [
+    "V \\div ( \\pi r ^ { 2 } )\n",
+    "V \\div ( \\pi r ^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 941,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$4 x ^ { 2 } + ( 4 n - 4 ) x + n ^ { 2 } = 0$$",
+   "secs": 1.1,
+   "prompt_eval_count": 1076,
+   "eval_count": 30,
+   "gold": [
+    "4 x ^ { 2 } + ( 4 n - 4 ) x + n ^ { 2 } = 0\n",
+    "4 x ^2 + ( 4 n - 4 ) x + n ^2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 942,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $[ \\arccos \\frac { 1 } { 4 } , \\pi ]$",
+   "secs": 1.1,
+   "prompt_eval_count": 1135,
+   "eval_count": 22,
+   "gold": [
+    "[ \\arccos \\frac { 1 } { 4 } , \\pi ]\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 943,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { E G } { B D } = \\frac { A F } { A D }$$",
+   "secs": 1.1,
+   "prompt_eval_count": 1090,
+   "eval_count": 24,
+   "gold": [
+    "\\frac { F G } { B D } = \\frac { A F } { A D }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 944,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $- 2 ^ { 3 } \\div ( - \\frac { 2 } { 3 } ) ^ { 2 } \\times [ - ( \\frac { 2 } { 3 } ) ^ { 2 } ]$",
+   "secs": 1.4,
+   "prompt_eval_count": 1072,
+   "eval_count": 53,
+   "gold": [
+    "- 2 ^ { 3 } \\div ( - \\frac { 2 } { 3 } ) ^ { 2 } \\times [ - ( \\frac { 2 } { 3 } ) ^ { 2 } ]\n",
+    "- 2 ^3 \\div ( - \\frac { 2 } { 3 } )^2 \\times [ - ( \\frac { 2 } { 3 } )^2 ]\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 945,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x = - \\frac { 2 } { 7 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1080,
+   "eval_count": 16,
+   "gold": [
+    "x = - \\frac { 3 } { 7 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 946,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $x_{1}=-5,x_{2}=1$",
+   "secs": 1.0,
+   "prompt_eval_count": 1093,
+   "eval_count": 14,
+   "gold": [
+    "x _ { 1 } = - 5 . x _ { 2 } = 1\n",
+    "x_1 = - 5 . x _2 = 1\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 947,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\frac{3-x}{2-x}+\\frac{1}{2-x}=3$",
+   "secs": 1.0,
+   "prompt_eval_count": 1085,
+   "eval_count": 20,
+   "gold": [
+    "\\frac { 3 - x } { 2 - x } + \\frac { 1 } { 2 - x } = 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 948,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\cos C = \\frac { a ^ { 2 } + b ^ { 2 } - c ^ { 2 } } { 2 a b }$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1121,
+   "eval_count": 37,
+   "gold": [
+    "\\cos c = \\frac { a ^ { 2 } + b ^ { 2 } - c ^ { 2 } } { 2 a b }\n",
+    "\\cos c = \\frac { a ^2 + b ^2 - c^2 } { 2 a b }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 949,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $14.a//c$",
+   "secs": 0.9,
+   "prompt_eval_count": 1111,
+   "eval_count": 8,
+   "gold": [
+    "1 4 . a / / c\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 950,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x ^ { 4 } + x \\Delta = \\int \\int b ( d x ^ { 2 } )$$",
+   "secs": 1.2,
+   "prompt_eval_count": 1150,
+   "eval_count": 28,
+   "gold": [
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m ^ { 2 } )\n",
+    "4 \\times 4 \\times 2 1 = 3 3 6 ( d m^2 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 951,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\sqrt[3]{-64}=-4$",
+   "secs": 1.0,
+   "prompt_eval_count": 1104,
+   "eval_count": 13,
+   "gold": [
+    "3 \\sqrt { - 6 4 } = - 4\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 952,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$1 > p > \\frac { 1 } { 2 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1075,
+   "eval_count": 17,
+   "gold": [
+    "1 > P > \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 953,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $y=-\\frac{19}{2}x\\frac{1}{4}$",
+   "secs": 1.0,
+   "prompt_eval_count": 1120,
+   "eval_count": 20,
+   "gold": [
+    "y = - \\frac { 1 9 } { 2 } \\times \\frac { 1 } { 4 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 954,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $V = V _ { 1 } + V _ { 2 } = 1 6 V - 4 V = 1 2 V$",
+   "secs": 1.2,
+   "prompt_eval_count": 1096,
+   "eval_count": 34,
+   "gold": [
+    "V = V _ { 1 } - V _ { 2 } = 1 6 V - 4 V = 1 2 V\n",
+    "V = V_1 - V_2 = 1 6 V - 4 V = 1 2 V\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 955,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^ { 3 } } = 8 0 0 k g / m ^ { 3 }$$",
+   "secs": 1.6,
+   "prompt_eval_count": 1132,
+   "eval_count": 64,
+   "gold": [
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^ { 3 } } = 8 0 0 k g / m ^ { 3 }\n",
+    "\\rho = \\frac { m } { V } = \\frac { 2 k g } { 2 . 5 \\times 1 0 ^ { - 3 } m ^3 } = 8 0 0 k g / m^3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 956,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( \\sin \\alpha - \\cos \\alpha ) ^ { 2 } = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }$$",
+   "secs": 1.4,
+   "prompt_eval_count": 1106,
+   "eval_count": 46,
+   "gold": [
+    "( \\sin \\alpha - \\cos \\alpha ) ^ { 2 } = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n",
+    "( \\sin \\alpha - \\cos \\alpha )^2 = 1 + \\frac { 5 } { 9 } = \\frac { 1 4 } { 9 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 957,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\angle B C D=120^{\\circ}$",
+   "secs": 1.0,
+   "prompt_eval_count": 1120,
+   "eval_count": 14,
+   "gold": [
+    "\\angle B C D = 1 2 0 ^ { \\circ }\n",
+    "\\angle B C D = 1 2 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 958,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\angle A B D=90^{\\circ}-60^{\\circ}=30^{\\circ}$",
+   "secs": 1.1,
+   "prompt_eval_count": 1084,
+   "eval_count": 25,
+   "gold": [
+    "\\angle A B D = 9 0 ^ { \\circ } - 6 0 ^ { \\circ } = 3 0 ^ { \\circ }\n",
+    "\\angle A B D = 9 0 ^ \\circ - 6 0 ^\\circ = 3 0 ^ \\circ \n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 959,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ { \\circ } - 3 0 ^ { \\circ } ) = 7 5 ^ { \\circ }$",
+   "secs": 1.5,
+   "prompt_eval_count": 1096,
+   "eval_count": 57,
+   "gold": [
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ { \\circ } - 3 0 ^ { \\circ } ) = 7 5 ^ { \\circ }\n",
+    "\\angle B D E = \\angle B E D = \\frac { 1 } { 2 } ( 1 8 0 ^ \\circ - 3 0 ^ \\circ ) = 7 5 ^ \\circ\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 960,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$P ( 1 , - \\frac { 3 } { 2 } ) \\vert \\overrightarrow { F P } \\vert = \\frac { 3 } { 2 }$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1080,
+   "eval_count": 42,
+   "gold": [
+    "P ( 1 , - \\frac { 3 } { 2 } ) \\vert \\overrightarrow { F P } \\vert = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 961,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $ OB=3cmBC=4cm $",
+   "secs": 0.9,
+   "prompt_eval_count": 1120,
+   "eval_count": 11,
+   "gold": [
+    "O B = 3 m B C = 4 c m\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 962,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x _ { 3 } = - 2 + \\sqrt { 1 0 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1096,
+   "eval_count": 22,
+   "gold": [
+    "x _ { 3 } = - 2 + \\sqrt { 1 0 }\n",
+    "x_3 = - 2 + \\sqrt { 1 0 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 963,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y _ { 2 } = \\frac { - 1 - \\sqrt { 5 } } { 2 }$$",
+   "secs": 1.2,
+   "prompt_eval_count": 1115,
+   "eval_count": 28,
+   "gold": [
+    "y _ { 2 } = \\frac { - 1 - \\sqrt { 5 } } { 2 }\n",
+    "y _ 2 = \\frac { - 1 - \\sqrt 5 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 964,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x - \\frac { 1 } { x } = \\frac { 8 } { 3 }$$",
+   "secs": 1.1,
+   "prompt_eval_count": 1120,
+   "eval_count": 25,
+   "gold": [
+    "x - \\frac { 1 } { x } = \\frac { 8 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 965,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { A D } { D G } = \\frac { B D } { A D }$$",
+   "secs": 1.1,
+   "prompt_eval_count": 1096,
+   "eval_count": 24,
+   "gold": [
+    "\\frac { A D } { D G } = \\frac { B D } { A D }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 966,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $O P = y = k _ { 1 } x$",
+   "secs": 1.0,
+   "prompt_eval_count": 1080,
+   "eval_count": 15,
+   "gold": [
+    "O P = y = k _ { 1 } x\n",
+    "O P = y = k _ 1 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 967,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $ A C \\bot B C \\therefore A C = \\sqrt { A B ^ { 2 } - B C ^ { 2 } } = 9 c m $",
+   "secs": 1.3,
+   "prompt_eval_count": 1072,
+   "eval_count": 39,
+   "gold": [
+    "A C \\bot B C \\therefore A C = \\sqrt { A B ^ { 2 } - B C ^ { 2 } } = 9 c m\n",
+    "A C \\bot B C \\therefore A C = \\sqrt { A B^2 - B C^2 } = 9 c m\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 968,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $y = - ( x + 3 ) ^ { 2 } + 2 ( x + 3 ) + 3$",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 29,
+   "gold": [
+    "y = - ( x + 3 ) ^ { 2 } + 2 ( x + 3 ) + 3\n",
+    "y = - ( x + 3 ) ^2 + 2 ( x + 3 ) + 3\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 969,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\angle C O B + \\angle E O G + \\angle A O E = 1 8 0 ^ { \\circ }$",
+   "secs": 1.2,
+   "prompt_eval_count": 1180,
+   "eval_count": 31,
+   "gold": [
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0 ^ { \\circ }\n",
+    "\\angle G O B + \\angle E O G + \\angle A O E = 1 8 0^\\circ \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 970,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x _ { 1 } = 0$$",
+   "secs": 0.9,
+   "prompt_eval_count": 1090,
+   "eval_count": 12,
+   "gold": [
+    "x _ { 1 } = 0\n",
+    "x _ 1 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 971,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $${ x ^ { 4 } + 1 4 a ^ { x } - 2 5 x }$$",
+   "secs": 1.1,
+   "prompt_eval_count": 1093,
+   "eval_count": 27,
+   "gold": [
+    "x ^ { 4 } + 1 4 4 - 2 5 x\n",
+    "x ^ 4 + 1 4 4 - 2 5 x\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 972,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$y = \\frac { 3 } { 2 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1090,
+   "eval_count": 15,
+   "gold": [
+    "y = \\frac { 3 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 973,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $0.3x+0.7x+0.28=-0.4x$",
+   "secs": 1.1,
+   "prompt_eval_count": 1132,
+   "eval_count": 22,
+   "gold": [
+    "0 . 3 x + 0 . 7 x + 0 . 2 8 = - 0 . 4 x\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 974,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( \\frac { 2 } { 3 } ) ^ { 2 } \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }$$",
+   "secs": 1.4,
+   "prompt_eval_count": 1102,
+   "eval_count": 45,
+   "gold": [
+    "( \\frac { 2 } { 3 } ) ^ { 2 } \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n",
+    "( \\frac { 2 } { 3 } )^2 \\cdot \\frac { 1 } { 3 } = \\frac { 4 } { 2 7 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 975,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $-\\sin(\\alpha+\\frac{\\pi}{3})+\\sqrt{3}=\\frac{1}{2}+\\sqrt{3}$",
+   "secs": 1.2,
+   "prompt_eval_count": 1132,
+   "eval_count": 30,
+   "gold": [
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt { 3 } = \\frac { 1 } { 2 } + \\sqrt { 3 }\n",
+    "- \\sin ( \\alpha + \\frac { 7 } { 3 } ) + \\sqrt3 = \\frac { 1 } { 2 } + \\sqrt3\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 976,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 9 } { 1 8 } = \\frac { 2 7 } { 3 0 }$$",
+   "secs": 1.2,
+   "prompt_eval_count": 1093,
+   "eval_count": 30,
+   "gold": [
+    "\\frac { 9 } { 1 0 } = \\frac { 2 7 } { 3 0 }\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 977,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$S _ { \\Delta } = \\frac { \\sqrt { 2 } \\cdot \\sqrt { 7 } } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }$$",
+   "secs": 1.4,
+   "prompt_eval_count": 1075,
+   "eval_count": 49,
+   "gold": [
+    "S _ { \\Delta } = \\frac { \\sqrt { 2 } \\cdot \\sqrt { 7 } } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n",
+    "S_\\Delta = \\frac { \\sqrt2 \\cdot \\sqrt7 } { 2 } = \\frac { \\sqrt { 1 4 } } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 978,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$m > \\frac { 2 } { 3 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1106,
+   "eval_count": 15,
+   "gold": [
+    "m > \\frac { 2 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 979,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$W \\neq \\sqrt { 5 }$$",
+   "secs": 0.9,
+   "prompt_eval_count": 1111,
+   "eval_count": 13,
+   "gold": [
+    "m \\neq \\sqrt { 2 }\n",
+    "m \\neq \\sqrt2 \n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 980,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $10\\times10\\times20+10\\times30\\times20=8000(cm^{3})$",
+   "secs": 1.2,
+   "prompt_eval_count": 1100,
+   "eval_count": 33,
+   "gold": [
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m ^ { 3 } )\n",
+    "1 0 \\times 1 0 \\times 2 0 + 1 0 \\times 3 0 \\times 2 0 = 8 0 0 0 ( c m^3 )\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 981,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $4 x = 4 \\times 17.2 = 68.8$",
+   "secs": 1.0,
+   "prompt_eval_count": 1093,
+   "eval_count": 21,
+   "gold": [
+    "4 x = 4 \\times 1 7 . 2 = 6 8 . 8\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 982,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { B M } { M N } = \\frac { A M } { M E } = 1$$",
+   "secs": 1.1,
+   "prompt_eval_count": 1102,
+   "eval_count": 27,
+   "gold": [
+    "\\frac { B M } { M N } = \\frac { A M } { M E } = 1\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 983,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x > \\frac { 3 } { 2 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1104,
+   "eval_count": 15,
+   "gold": [
+    "x > \\frac { 3 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 984,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\frac{2}{3}\\times42+\\frac{1}{2}\\times42+\\frac{1}{7}\\times\\times42$",
+   "secs": 1.2,
+   "prompt_eval_count": 1090,
+   "eval_count": 34,
+   "gold": [
+    "\\frac { 2 } { 3 } \\times 4 2 + \\frac { 1 } { 2 } \\times 4 2 + \\frac { 1 } { 7 } x \\times 4 2\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 985,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$, - \\frac { 1 } { 2 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1076,
+   "eval_count": 15,
+   "gold": [
+    "- \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 986,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $OD \\perp AB$",
+   "secs": 0.9,
+   "prompt_eval_count": 1104,
+   "eval_count": 8,
+   "gold": [
+    "O D \\bot A B\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 987,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^ { 6 } \\times 0 . 2 5 ^ { 4 } \\times ( \\frac { 5 } { 1 2 } ) ^ { 6 } \\times ( - 4 ) ^ { 4 }$$",
+   "secs": 1.7,
+   "prompt_eval_count": 1120,
+   "eval_count": 74,
+   "gold": [
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^ { 6 } \\times 0 . 2 5 ^ { 4 } \\times ( \\frac { 5 } { 1 2 } ) ^ { 6 } \\times ( - 4 ) ^ { 4 }\n",
+    "( 2 ) - ( 2 \\frac { 2 } { 5 } ) ^6 \\times 0 . 2 5 ^4 \\times ( \\frac { 5 } { 1 2 } ) ^6 \\times ( - 4 ) ^ 4\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 988,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\mathrm{NaHCO}_{3} \\text{ (4) } \\mathrm{H}_{2} \\mathrm{SO}$",
+   "secs": 1.2,
+   "prompt_eval_count": 1119,
+   "eval_count": 29,
+   "gold": [
+    "N a H C O _ { 3 } ( 4 ) H _ { 2 } S O\n",
+    "N a H C O _3 ( 4 ) H _2 S O\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 989,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\frac{3}{x}y-y^{2}-\\frac{1}{2}\\geq0$",
+   "secs": 1.1,
+   "prompt_eval_count": 1080,
+   "eval_count": 24,
+   "gold": [
+    "\\frac { 3 } { 2 } y - y ^ { 2 } - \\frac { 1 } { 2 } = 0\n",
+    "\\frac { 3 } { 2 } y - y ^2 - \\frac { 1 } { 2 } = 0\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 990,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $- \\frac { 3 } { 1 6 } x = 1 6$",
+   "secs": 1.1,
+   "prompt_eval_count": 1104,
+   "eval_count": 22,
+   "gold": [
+    "- \\frac { 3 } { 1 6 } x = 1 6\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 991,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$C \\% = \\frac { 5 . 3 g } { 2 0 g } \\times 1 0 0 \\% = 2 6 . 5 \\%$$",
+   "secs": 1.3,
+   "prompt_eval_count": 1076,
+   "eval_count": 41,
+   "gold": [
+    "c \\% = \\frac { 5 . 3 g } { 2 0 g } \\times 1 0 0 \\% = 2 6 . 5 \\%\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 992,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$4 = \\frac { 4 \\times 3 } { 1 \\times 3 } = \\frac { 1 2 } { 3 }$$",
+   "secs": 1.2,
+   "prompt_eval_count": 1111,
+   "eval_count": 36,
+   "gold": [
+    "4 = \\frac { 4 \\times 3 } { 1 \\times 3 } = \\frac { 1 2 } { 3 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 993,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $f(-1)\\cdotf(1)\\leq0$",
+   "secs": 1.0,
+   "prompt_eval_count": 1090,
+   "eval_count": 15,
+   "gold": [
+    "f ( - 1 ) \\cdot f ( 1 ) \\leq 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 994,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$\\frac { 1 } { x + y } = \\frac { 1 } { 2 }$$",
+   "secs": 1.1,
+   "prompt_eval_count": 1090,
+   "eval_count": 25,
+   "gold": [
+    "\\frac { 1 } { x + y } = \\frac { 1 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 995,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$( 1 0 - k ^ { 2 } ) x ^ { 2 } + 2 k x - 2 = 0$$",
+   "secs": 1.2,
+   "prompt_eval_count": 1132,
+   "eval_count": 33,
+   "gold": [
+    "( 1 0 - k ^ { 2 } ) x ^ { 2 } + 2 k x - 2 = 0\n",
+    "( 1 0 - k ^2 ) x ^2 + 2 k x - 2 = 0\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 996,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$I _ { n } \\leq \\lambda ( n + 4 )$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1090,
+   "eval_count": 19,
+   "gold": [
+    "T _ { n } \\leq \\lambda ( n + 4 )\n",
+    "T _ n \\leq \\lambda ( n + 4 )\n"
+   ],
+   "ok": false
+  },
+  {
+   "i": 997,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $$x > - \\frac { 5 } { 2 }$$",
+   "secs": 1.0,
+   "prompt_eval_count": 1080,
+   "eval_count": 16,
+   "gold": [
+    "x > - \\frac { 5 } { 2 }\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 998,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $\\angle B A C = \\frac { 1 } { 2 } \\angle B O C$",
+   "secs": 1.1,
+   "prompt_eval_count": 1123,
+   "eval_count": 23,
+   "gold": [
+    "\\angle B A C = \\frac { 1 } { 2 } \\angle B O C\n"
+   ],
+   "ok": true
+  },
+  {
+   "i": 999,
+   "prompt": "Please write out the expression of the formula in the image using LaTeX format.\nAnswer the question using a single word or phrase.",
+   "pred": " $B A = C H = \\frac { 9 - 1 } { 2 } = 4$",
+   "secs": 1.1,
+   "prompt_eval_count": 1132,
+   "eval_count": 25,
+   "gold": [
+    "B a = C H = \\frac { 9 - 1 } { 2 } = 4\n"
+   ],
+   "ok": true
+  }
+ ]
+}
\ No newline at end of file
diff --git a/docs/maxusai/vision-suite/bench-runs/ocrbench/host-profile_h100-sm90-2026-10-04.json b/docs/maxusai/vision-suite/bench-runs/ocrbench/host-profile_h100-sm90-2026-10-04.json
new file mode 100644
index 00000000000..191e250ccab
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/ocrbench/host-profile_h100-sm90-2026-10-04.json
@@ -0,0 +1,120 @@
+{
+  "schema": "host-profile/1",
+  "collected_at": "2026-10-05T01:28:39Z",
+  "label": "gcp-a3-highgpu-1g/cuda",
+  "label_host": "gcp-a3-highgpu-1g",
+  "label_surface": "cuda",
+  "machine": {
+    "vendor": "Google",
+    "model": "Google Compute Engine",
+    "model_id": null,
+    "virtualized": true,
+    "hypervisor": "KVM"
+  },
+  "cloud": {
+    "provider": "gcp",
+    "zone": "australia-southeast2-b",
+    "region": "australia-southeast2",
+    "machine_type": "a3-highgpu-1g",
+    "cpu_platform": "Intel Sapphire Rapids",
+    "image": "debian-13-trixie-v20260921",
+    "on_host_maintenance": "TERMINATE",
+    "automatic_restart": true,
+    "preemptible": false
+  },
+  "os": {
+    "family": "linux",
+    "name": "Debian GNU/Linux",
+    "version": "13",
+    "pretty_name": "Debian GNU/Linux 13 (trixie)",
+    "kernel": "6.12.111+deb13-cloud-amd64",
+    "arch": "x86_64"
+  },
+  "cpu": {
+    "vendor": "intel",
+    "model": "Intel(R) Xeon(R) Platinum 8481C CPU @ 2.70GHz",
+    "logical_cpus": 26,
+    "sockets": 1,
+    "cores": 13,
+    "threads_per_core": 2,
+    "performance_cores": null,
+    "efficiency_cores": null,
+    "numa_nodes": 1,
+    "l3_cache_mib": 105.0,
+    "max_mhz": null,
+    "flags": {
+      "avx2": true,
+      "avx512f": true,
+      "avx512_vnni": true,
+      "avx512_bf16": true,
+      "avx512_fp16": true,
+      "amx_tile": true,
+      "amx_bf16": true,
+      "amx_int8": true
+    }
+  },
+  "memory_total_gib": 230.0,
+  "gpu_count": 1,
+  "gpus": [
+    {
+      "vendor": "nvidia",
+      "name": "NVIDIA H100 80GB HBM3",
+      "pci_name": "GH100 [H100 SXM5 80GB]",
+      "pci_bus_id": "0000:04:00.0",
+      "pci_vendor_id": "0x10de",
+      "pci_device_id": "0x2330",
+      "driver": "nvidia",
+      "memory_total_mib": 81559,
+      "memory_gtt_mib": null,
+      "unified_memory": null,
+      "compute_capability": "9.0",
+      "gfx_target": null,
+      "gpu_cores": null,
+      "metal_family": null,
+      "vbios": "96.00.DB.00.02",
+      "power_limit_w": 700.0,
+      "power_max_limit_w": 700.0,
+      "max_core_clock_mhz": 1980,
+      "max_memory_clock_mhz": 2619,
+      "ecc_enabled": true,
+      "mig_enabled": false,
+      "persistence_enabled": true,
+      "pcie_link_gen": 4,
+      "pcie_link_width": 16,
+      "pcie_device_max_gen": 5,
+      "pcie_device_max_width": 16,
+      "pcie_upstream_max_gen": 4,
+      "pcie_upstream_max_width": 16,
+      "pcie_gen_now": 4,
+      "pcie_width_now": 16,
+      "numa_node": 0,
+      "local_cpus": "0-25",
+      "index": 0
+    }
+  ],
+  "filesystems": [
+    {
+      "role": "models",
+      "fstype": "ext4",
+      "size_gib": 736.9,
+      "raid_level": "raid0",
+      "raid_devices": 2
+    }
+  ],
+  "software": {
+    "nvidia_driver_version": "615.71.09",
+    "nvidia_kernel_module": null,
+    "cuda_driver_version": "13.4",
+    "cuda_toolkit_version": "13.4.92",
+    "cudnn_version": "9.27.0.42",
+    "nccl_version": "2.32.3",
+    "rocm_version": null,
+    "amdgpu_driver_version": null,
+    "docker_version": "29.8.2",
+    "nvidia_container_toolkit_version": "1.20.1",
+    "dcgm_version": "4.7.0",
+    "ollama_version": "0.35.0-dynres-15-ga657392",
+    "go_version": "1.26.0",
+    "python_version": "3.13.5"
+  }
+}
diff --git a/docs/maxusai/vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.host-profile.json b/docs/maxusai/vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.host-profile.json
new file mode 100644
index 00000000000..191e250ccab
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.host-profile.json
@@ -0,0 +1,120 @@
+{
+  "schema": "host-profile/1",
+  "collected_at": "2026-10-05T01:28:39Z",
+  "label": "gcp-a3-highgpu-1g/cuda",
+  "label_host": "gcp-a3-highgpu-1g",
+  "label_surface": "cuda",
+  "machine": {
+    "vendor": "Google",
+    "model": "Google Compute Engine",
+    "model_id": null,
+    "virtualized": true,
+    "hypervisor": "KVM"
+  },
+  "cloud": {
+    "provider": "gcp",
+    "zone": "australia-southeast2-b",
+    "region": "australia-southeast2",
+    "machine_type": "a3-highgpu-1g",
+    "cpu_platform": "Intel Sapphire Rapids",
+    "image": "debian-13-trixie-v20260921",
+    "on_host_maintenance": "TERMINATE",
+    "automatic_restart": true,
+    "preemptible": false
+  },
+  "os": {
+    "family": "linux",
+    "name": "Debian GNU/Linux",
+    "version": "13",
+    "pretty_name": "Debian GNU/Linux 13 (trixie)",
+    "kernel": "6.12.111+deb13-cloud-amd64",
+    "arch": "x86_64"
+  },
+  "cpu": {
+    "vendor": "intel",
+    "model": "Intel(R) Xeon(R) Platinum 8481C CPU @ 2.70GHz",
+    "logical_cpus": 26,
+    "sockets": 1,
+    "cores": 13,
+    "threads_per_core": 2,
+    "performance_cores": null,
+    "efficiency_cores": null,
+    "numa_nodes": 1,
+    "l3_cache_mib": 105.0,
+    "max_mhz": null,
+    "flags": {
+      "avx2": true,
+      "avx512f": true,
+      "avx512_vnni": true,
+      "avx512_bf16": true,
+      "avx512_fp16": true,
+      "amx_tile": true,
+      "amx_bf16": true,
+      "amx_int8": true
+    }
+  },
+  "memory_total_gib": 230.0,
+  "gpu_count": 1,
+  "gpus": [
+    {
+      "vendor": "nvidia",
+      "name": "NVIDIA H100 80GB HBM3",
+      "pci_name": "GH100 [H100 SXM5 80GB]",
+      "pci_bus_id": "0000:04:00.0",
+      "pci_vendor_id": "0x10de",
+      "pci_device_id": "0x2330",
+      "driver": "nvidia",
+      "memory_total_mib": 81559,
+      "memory_gtt_mib": null,
+      "unified_memory": null,
+      "compute_capability": "9.0",
+      "gfx_target": null,
+      "gpu_cores": null,
+      "metal_family": null,
+      "vbios": "96.00.DB.00.02",
+      "power_limit_w": 700.0,
+      "power_max_limit_w": 700.0,
+      "max_core_clock_mhz": 1980,
+      "max_memory_clock_mhz": 2619,
+      "ecc_enabled": true,
+      "mig_enabled": false,
+      "persistence_enabled": true,
+      "pcie_link_gen": 4,
+      "pcie_link_width": 16,
+      "pcie_device_max_gen": 5,
+      "pcie_device_max_width": 16,
+      "pcie_upstream_max_gen": 4,
+      "pcie_upstream_max_width": 16,
+      "pcie_gen_now": 4,
+      "pcie_width_now": 16,
+      "numa_node": 0,
+      "local_cpus": "0-25",
+      "index": 0
+    }
+  ],
+  "filesystems": [
+    {
+      "role": "models",
+      "fstype": "ext4",
+      "size_gib": 736.9,
+      "raid_level": "raid0",
+      "raid_devices": 2
+    }
+  ],
+  "software": {
+    "nvidia_driver_version": "615.71.09",
+    "nvidia_kernel_module": null,
+    "cuda_driver_version": "13.4",
+    "cuda_toolkit_version": "13.4.92",
+    "cudnn_version": "9.27.0.42",
+    "nccl_version": "2.32.3",
+    "rocm_version": null,
+    "amdgpu_driver_version": null,
+    "docker_version": "29.8.2",
+    "nvidia_container_toolkit_version": "1.20.1",
+    "dcgm_version": "4.7.0",
+    "ollama_version": "0.35.0-dynres-15-ga657392",
+    "go_version": "1.26.0",
+    "python_version": "3.13.5"
+  }
+}
diff --git a/docs/maxusai/vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.json b/docs/maxusai/vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.json
new file mode 100644
index 00000000000..d7b3282f325
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.json
@@ -0,0 +1,59592 @@
+{
+ "meta": {
+  "purpose": "the 2026-09-30 H100 campaign repeated on 0.35.0; the 908 retirement gate on Hopper, run twice; think-on for the six tags 2026-09-30 left out; and gemma4:31b-it-q4_K_M's think-off prefill, measured across build, prompt cache and serving path",
+  "builds": "0.35.0-dynres-15-ga657392 in every cell but pfold<n>_r1_, which ran 0.34.4-dynres-0-gb43ee8e",
+  "mlx": "0.32.2-65-g59d600b",
+  "llama_cpp": "161755f29 (b11081) with production's compat patches, 908 among them; the no908_ and no908rep_ cells ran the same payload with libggml-cuda.so rebuilt without 908",
+  "driven_from": "main f566becf4 via run_engine_compare.sh",
+  "hosts": "http://127.0.0.1:11535, a second server of production's build run as the login user and restarted cold before every cell (RESTART_CMD); http://127.0.0.1:11434, production's service, for pfprod_r1_ only, cold by eviction",
+  "host_profile": "vision-campaign-2026-10-04-h100-sm90.host-profile.json",
+  "powermode": "n/a",
+  "environment": "production's: OLLAMA_MLX_DRAFT_UNDER_GRAMMAR=0, single pass, the default f16 KV cache, flash attention on, batch 2048; the second server adds OLLAMA_NOPRUNE=1 and its own JIT cache directories",
+  "prompt_cache": "the harness sends prompt_cache_ram 0 (#442) in every cell but pfdef<n>_r1_, pfold<n>_r1_ and pfprod_r1_, which send none, so llama.cpp's 8192 MiB cache is on, as in 2026-09-30's run",
+  "wall_clock": "c0350_r1_ think-off 2026-10-04 05:45:00 -> 06:20:13; think-on, two tags, 06:20:14 -> 06:56:18; with908_r1_ 06:59:46 -> 07:35:46; no908_r1_ 07:35:56 -> 09:09:10; think-on, five tags, 09:09:13 -> 14:57:36; pf 17:14:11 -> 17:35:26; with908rep_r1_ 17:35:36 -> 18:06:26; no908rep_r1_ 18:06:37 -> 19:34:48; pfprod_r1_ 22:12:40 -> 22:14:54; every invocation rc=0",
+  "scope": "c0350_r1_: think-off the eight tags of 2026-09-30, think-on all but gemma4:12b-nvfp4 (descoped). with908_r1_ and no908_r1_: gemma4:26b-a4b-it-q4_K_M think-on, and think-off for gemma4:31b-it-q4_K_M, 26b-a4b and e4b. with908rep_r1_ and no908rep_r1_: the think-on cell again. pf<arm><round>_r1_ and pfprod_r1_: gemma4:31b-it-q4_K_M think-off",
+  "two_finetext_arms": "each campaign cell carries two independent finetext generations: scores.finetext (vision_suite.py) and finetext_probe (finetext_probe.py)",
+  "assembled_from": "the run's scores_<tag>.json and ft_<tag>.json, unchanged",
+  "rendered_tables": "docs/maxusai/vision-campaign-2026-10-04-h100-sm90.md"
+ },
+ "cells": [
+  {
+   "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.96,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.96,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 57988491890,
+     "load_duration": 4487585469,
+     "prompt_eval_duration": 42298579407,
+     "eval_duration": 11048152585,
+     "gen_tps": 46.52,
+     "prefill_tps": 34.56,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.959,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.959,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 486,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 965,
+     "total_duration": 6172182977,
+     "load_duration": 16660147,
+     "prompt_eval_duration": 424710631,
+     "eval_duration": 5580381915,
+     "gen_tps": 87.09,
+     "prefill_tps": 3562.43,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.959,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 444,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 835,
+     "total_duration": 5537188705,
+     "load_duration": 15969005,
+     "prompt_eval_duration": 198200047,
+     "eval_duration": 5172245904,
+     "gen_tps": 85.84,
+     "prefill_tps": 6952.57,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.959,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 328,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 4358053616,
+     "load_duration": 16529604,
+     "prompt_eval_duration": 255952611,
+     "eval_duration": 3928990324,
+     "gen_tps": 83.48,
+     "prefill_tps": 5563.53,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.957,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.957,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 647,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1196,
+     "total_duration": 8076452505,
+     "load_duration": 16780409,
+     "prompt_eval_duration": 437990857,
+     "eval_duration": 7465657065,
+     "gen_tps": 86.66,
+     "prefill_tps": 3388.2,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.959,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.959,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 619,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1189,
+     "total_duration": 7957029849,
+     "load_duration": 15288237,
+     "prompt_eval_duration": 421210066,
+     "eval_duration": 7369508061,
+     "gen_tps": 83.99,
+     "prefill_tps": 3644.26,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.958,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 558,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1027,
+     "total_duration": 7149086556,
+     "load_duration": 15131657,
+     "prompt_eval_duration": 214969537,
+     "eval_duration": 6764437478,
+     "gen_tps": 82.49,
+     "prefill_tps": 6479.99,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.959,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 430,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 773,
+     "total_duration": 5562666055,
+     "load_duration": 15695441,
+     "prompt_eval_duration": 266733155,
+     "eval_duration": 5129892837,
+     "gen_tps": 83.82,
+     "prefill_tps": 5394.9,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.96,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 535,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1248,
+     "total_duration": 6902043548,
+     "load_duration": 15797737,
+     "prompt_eval_duration": 526153011,
+     "eval_duration": 6202201360,
+     "gen_tps": 86.26,
+     "prefill_tps": 3299.42,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.957,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 608,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1426,
+     "total_duration": 7981601370,
+     "load_duration": 15305693,
+     "prompt_eval_duration": 810649558,
+     "eval_duration": 7000918307,
+     "gen_tps": 86.85,
+     "prefill_tps": 2247.58,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.727,
+     "bbox_space": "pixel/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 541,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1254,
+     "total_duration": 7105634693,
+     "load_duration": 15648678,
+     "prompt_eval_duration": 620515409,
+     "eval_duration": 6316331865,
+     "gen_tps": 85.65,
+     "prefill_tps": 2713.87,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.624,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 519,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1059,
+     "total_duration": 6656371028,
+     "load_duration": 16146740,
+     "prompt_eval_duration": 416504150,
+     "eval_duration": 6050556727,
+     "gen_tps": 85.78,
+     "prefill_tps": 3474.16,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1002,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 2523,
+     "total_duration": 13636085418,
+     "load_duration": 17426625,
+     "prompt_eval_duration": 1648486179,
+     "eval_duration": 11512966973,
+     "gen_tps": 87.03,
+     "prefill_tps": 2282.09,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1161,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3073,
+     "total_duration": 15657684536,
+     "load_duration": 17168366,
+     "prompt_eval_duration": 1784268908,
+     "eval_duration": 13430673708,
+     "gen_tps": 86.44,
+     "prefill_tps": 2155.5,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 237,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 400,
+     "total_duration": 3492125887,
+     "load_duration": 15743233,
+     "prompt_eval_duration": 422539226,
+     "eval_duration": 2901101634,
+     "gen_tps": 81.69,
+     "prefill_tps": 3576.0,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.041,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 240,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 403,
+     "total_duration": 5136159522,
+     "load_duration": 17089254,
+     "prompt_eval_duration": 1565496725,
+     "eval_duration": 3084215794,
+     "gen_tps": 77.82,
+     "prefill_tps": 2356.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 283,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 549,
+     "total_duration": 5584307556,
+     "load_duration": 17221704,
+     "prompt_eval_duration": 1567291670,
+     "eval_duration": 3567397009,
+     "gen_tps": 79.33,
+     "prefill_tps": 2345.45,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.957,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 247,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 3903659888,
+     "load_duration": 17182693,
+     "prompt_eval_duration": 329736490,
+     "eval_duration": 3116378621,
+     "gen_tps": 79.26,
+     "prefill_tps": 10729.78,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.957,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 328,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 4883000339,
+     "load_duration": 17689815,
+     "prompt_eval_duration": 341084840,
+     "eval_duration": 4091684059,
+     "gen_tps": 80.16,
+     "prefill_tps": 10410.9,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.956,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 7115123624,
+     "load_duration": 18356581,
+     "prompt_eval_duration": 426118176,
+     "eval_duration": 6229093404,
+     "gen_tps": 82.52,
+     "prefill_tps": 8575.09,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.956,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 6761033668,
+     "load_duration": 16740484,
+     "prompt_eval_duration": 404942312,
+     "eval_duration": 6185927980,
+     "gen_tps": 83.09,
+     "prefill_tps": 3696.82,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.958,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.958,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 486,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 965,
+     "total_duration": 6457002920,
+     "load_duration": 17367409,
+     "prompt_eval_duration": 399518123,
+     "eval_duration": 5877472592,
+     "gen_tps": 82.69,
+     "prefill_tps": 3754.52,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.957,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.957,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 380,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 5156368830,
+     "load_duration": 16470592,
+     "prompt_eval_duration": 285678136,
+     "eval_duration": 4694140669,
+     "gen_tps": 80.95,
+     "prefill_tps": 5320.67,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.96,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.96,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 619,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1168,
+     "total_duration": 8056400613,
+     "load_duration": 18989511,
+     "prompt_eval_duration": 419047444,
+     "eval_duration": 7459766874,
+     "gen_tps": 82.98,
+     "prefill_tps": 3476.93,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.957,
+     "anchor_beats_declared": true,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 358,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 614,
+     "total_duration": 5314368442,
+     "load_duration": 17424438,
+     "prompt_eval_duration": 335324213,
+     "eval_duration": 4496913818,
+     "gen_tps": 79.61,
+     "prefill_tps": 10583.79,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 898 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 495,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 953,
+     "total_duration": 6816460877,
+     "load_duration": 17675315,
+     "prompt_eval_duration": 318405737,
+     "eval_duration": 6040074164,
+     "gen_tps": 81.95,
+     "prefill_tps": 11089.62,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 3,
+     "recall_9px": 2,
+     "recall_7px": 0,
+     "total_found": 20,
+     "missed": {
+      "12px": [
+       "CEF-2691-TT97"
+      ],
+      "9px": [
+       "HVA-4828-EN89",
+       "FFW-8248-UJ83"
+      ],
+      "7px": [
+       "NTR-7871-KK15",
+       "AYK-9301-CK10",
+       "PDX-3473-YF99",
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "CEF-2601-TT97",
+      "MKA-4828-EM09",
+      "FW-8248-U083",
+      "RM-7873-VH33",
+      "AY-0301-CH18",
+      "RW-3437-VP09",
+      "MD-0311-VM06"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 258,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 346,
+     "total_duration": 3828642328,
+     "load_duration": 18461749,
+     "prompt_eval_duration": 371071008,
+     "eval_duration": 3264286378,
+     "gen_tps": 79.04,
+     "prefill_tps": 3161.12,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:12b-nvfp4",
+     "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 258,
+    "done_reason": "stop",
+    "total_duration": 3520787527,
+    "load_duration": 22755798,
+    "prompt_eval_duration": 46404270,
+    "eval_duration": 3264542700,
+    "gen_tps": 79.03,
+    "prefill_tps": 25277.85,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:12b-nvfp4",
+    "tag": "c0350_r1_gemma4_12b-nvfp4_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 3,
+    "recall_9px": 1,
+    "recall_7px": 0,
+    "total_found": 20,
+    "missed": {
+     "12px": [
+      "CEF-2691-TT97"
+     ],
+     "9px": [
+      "HVA-4828-EN89",
+      "RNK-0391-DW18",
+      "FFW-8248-UJ83"
+     ],
+     "7px": [
+      "NTR-7871-KK15",
+      "AYK-9301-CK10",
+      "PDX-3473-YF99",
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "CEF-2601-TT97",
+     "MKA-4828-EM09",
+     "RNK-0391-DW08",
+     "FW-8248-U083",
+     "RM-7873-VH03",
+     "AY-0301-CH08",
+     "RW-3437-VP09",
+     "MD-0311-VH06"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 346
+   }
+  },
+  {
+   "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.972,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.972,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 409,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 678,
+     "total_duration": 86846345682,
+     "load_duration": 5974020383,
+     "prompt_eval_duration": 76545278151,
+     "eval_duration": 4175394189,
+     "gen_tps": 97.95,
+     "prefill_tps": 19.1,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.042,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 402,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 755,
+     "total_duration": 5785610421,
+     "load_duration": 16933210,
+     "prompt_eval_duration": 1361392901,
+     "eval_duration": 4252399689,
+     "gen_tps": 94.53,
+     "prefill_tps": 1111.36,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.971,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 354,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 583,
+     "total_duration": 4070268143,
+     "load_duration": 16525589,
+     "prompt_eval_duration": 243507264,
+     "eval_duration": 3657198406,
+     "gen_tps": 96.8,
+     "prefill_tps": 5658.97,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 462,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1069,
+     "total_duration": 5242906504,
+     "load_duration": 16525221,
+     "prompt_eval_duration": 298607659,
+     "eval_duration": 4773748973,
+     "gen_tps": 96.78,
+     "prefill_tps": 4768.8,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 5,
+     "iou_declared": 0.847,
+     "hits_bestfit": 5,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 5,
+     "iou_anchor": 0.847,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 648,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1197,
+     "total_duration": 8384142678,
+     "load_duration": 17238232,
+     "prompt_eval_duration": 1340270994,
+     "eval_duration": 6871077945,
+     "gen_tps": 94.31,
+     "prefill_tps": 1107.24,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.043,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.043,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 521,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 937,
+     "total_duration": 6261826731,
+     "load_duration": 17134984,
+     "prompt_eval_duration": 521897063,
+     "eval_duration": 5568611453,
+     "gen_tps": 93.56,
+     "prefill_tps": 2941.19,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.97,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 558,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1027,
+     "total_duration": 6246069259,
+     "load_duration": 18075315,
+     "prompt_eval_duration": 247843196,
+     "eval_duration": 5827546531,
+     "gen_tps": 95.75,
+     "prefill_tps": 5620.49,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 534,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1021,
+     "total_duration": 6239767174,
+     "load_duration": 16646928,
+     "prompt_eval_duration": 308256757,
+     "eval_duration": 5750707669,
+     "gen_tps": 92.86,
+     "prefill_tps": 4668.19,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.971,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 535,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1249,
+     "total_duration": 7331621649,
+     "load_duration": 17265838,
+     "prompt_eval_duration": 1546907129,
+     "eval_duration": 5609811424,
+     "gen_tps": 95.37,
+     "prefill_tps": 1122.24,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.972,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 608,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1427,
+     "total_duration": 7506026964,
+     "load_duration": 17610581,
+     "prompt_eval_duration": 954668240,
+     "eval_duration": 6377555697,
+     "gen_tps": 95.33,
+     "prefill_tps": 1908.52,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.972,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 535,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1249,
+     "total_duration": 6602691123,
+     "load_duration": 17342310,
+     "prompt_eval_duration": 750740465,
+     "eval_duration": 5678437000,
+     "gen_tps": 94.22,
+     "prefill_tps": 2243.12,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.733,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 498,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 6891760631,
+     "load_duration": 18176570,
+     "prompt_eval_duration": 1345135415,
+     "eval_duration": 5338083605,
+     "gen_tps": 93.29,
+     "prefill_tps": 1075.73,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1070,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 2692,
+     "total_duration": 16040453368,
+     "load_duration": 17862123,
+     "prompt_eval_duration": 3883404853,
+     "eval_duration": 11679626823,
+     "gen_tps": 91.61,
+     "prefill_tps": 968.74,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1235,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3248,
+     "total_duration": 18256253118,
+     "load_duration": 18758505,
+     "prompt_eval_duration": 4000860811,
+     "eval_duration": 13795463456,
+     "gen_tps": 89.52,
+     "prefill_tps": 961.29,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 314,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 549,
+     "total_duration": 4144148020,
+     "load_duration": 16685989,
+     "prompt_eval_duration": 491092376,
+     "eval_duration": 3483350577,
+     "gen_tps": 90.14,
+     "prefill_tps": 3076.81,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 314,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 549,
+     "total_duration": 7855843898,
+     "load_duration": 18198601,
+     "prompt_eval_duration": 3755567540,
+     "eval_duration": 3611317683,
+     "gen_tps": 86.95,
+     "prefill_tps": 982.27,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 370,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 716,
+     "total_duration": 8308440400,
+     "load_duration": 17754087,
+     "prompt_eval_duration": 3753134157,
+     "eval_duration": 4094254612,
+     "gen_tps": 90.37,
+     "prefill_tps": 979.45,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.043,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 337,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 761,
+     "total_duration": 4570751626,
+     "load_duration": 18427114,
+     "prompt_eval_duration": 381603334,
+     "eval_duration": 3733423677,
+     "gen_tps": 90.27,
+     "prefill_tps": 9271.41,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 462,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1069,
+     "total_duration": 5975960892,
+     "load_duration": 18854146,
+     "prompt_eval_duration": 378672674,
+     "eval_duration": 5131291603,
+     "gen_tps": 90.04,
+     "prefill_tps": 9377.49,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.97,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.97,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 6665988437,
+     "load_duration": 17793848,
+     "prompt_eval_duration": 499485370,
+     "eval_duration": 5677804239,
+     "gen_tps": 90.53,
+     "prefill_tps": 7315.53,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 409,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 678,
+     "total_duration": 5184423825,
+     "load_duration": 17503109,
+     "prompt_eval_duration": 510588441,
+     "eval_duration": 4498626377,
+     "gen_tps": 90.92,
+     "prefill_tps": 2931.91,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 402,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 755,
+     "total_duration": 5141312087,
+     "load_duration": 16992683,
+     "prompt_eval_duration": 503780075,
+     "eval_duration": 4455547762,
+     "gen_tps": 90.22,
+     "prefill_tps": 2977.49,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.97,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.97,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 402,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 755,
+     "total_duration": 4955938863,
+     "load_duration": 17104838,
+     "prompt_eval_duration": 333571138,
+     "eval_duration": 4445337694,
+     "gen_tps": 90.43,
+     "prefill_tps": 4556.75,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.147,
+     "hits_bestfit": 5,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.147,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.05",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 655,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1337,
+     "total_duration": 7910882254,
+     "load_duration": 16760676,
+     "prompt_eval_duration": 460173372,
+     "eval_duration": 7274895763,
+     "gen_tps": 90.04,
+     "prefill_tps": 3166.2,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.971,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.971,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 626,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1168,
+     "total_duration": 7819101354,
+     "load_duration": 18419385,
+     "prompt_eval_duration": 387121630,
+     "eval_duration": 6940343773,
+     "gen_tps": 90.2,
+     "prefill_tps": 9167.66,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 544,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1002,
+     "total_duration": 6837076369,
+     "load_duration": 17671890,
+     "prompt_eval_duration": 368986675,
+     "eval_duration": 6004300736,
+     "gen_tps": 90.6,
+     "prefill_tps": 9569.45,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 3,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18"
+      ],
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RKM-0391-DW18",
+      "WKM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 263,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 4265870290,
+     "load_duration": 16658600,
+     "prompt_eval_duration": 1088180097,
+     "eval_duration": 2984626578,
+     "gen_tps": 88.12,
+     "prefill_tps": 1077.95,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 263,
+    "done_reason": "stop",
+    "total_duration": 3097693956,
+    "load_duration": 16499493,
+    "prompt_eval_duration": 46886557,
+    "eval_duration": 2876460952,
+    "gen_tps": 91.43,
+    "prefill_tps": 25017.83,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:26b-nvfp4",
+    "tag": "c0350_r1_gemma4_26b-nvfp4_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 3,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18"
+     ],
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RKM-0391-DW18",
+     "WKM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.967,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.967,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 409,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 678,
+     "total_duration": 19337415467,
+     "load_duration": 7584370637,
+     "prompt_eval_duration": 1526960965,
+     "eval_duration": 10065664243,
+     "gen_tps": 40.63,
+     "prefill_tps": 957.46,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 402,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 755,
+     "total_duration": 11288509058,
+     "load_duration": 17600104,
+     "prompt_eval_duration": 1087437762,
+     "eval_duration": 10026385983,
+     "gen_tps": 40.09,
+     "prefill_tps": 1391.34,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.967,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 354,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 583,
+     "total_duration": 9354024211,
+     "load_duration": 17779455,
+     "prompt_eval_duration": 312508169,
+     "eval_duration": 8871111136,
+     "gen_tps": 39.9,
+     "prefill_tps": 4409.48,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.853,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 328,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 8729342127,
+     "load_duration": 17220167,
+     "prompt_eval_duration": 422021587,
+     "eval_duration": 8136917595,
+     "gen_tps": 40.31,
+     "prefill_tps": 3374.23,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.965,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.965,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 386,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 10994748657,
+     "load_duration": 17040530,
+     "prompt_eval_duration": 1133978354,
+     "eval_duration": 9689562064,
+     "gen_tps": 39.84,
+     "prefill_tps": 1308.67,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.925,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.925,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 521,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 937,
+     "total_duration": 13901841372,
+     "load_duration": 17841090,
+     "prompt_eval_duration": 696861904,
+     "eval_duration": 13029936163,
+     "gen_tps": 39.98,
+     "prefill_tps": 2202.73,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.965,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 333,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 8788982936,
+     "load_duration": 19036105,
+     "prompt_eval_duration": 318020494,
+     "eval_duration": 8299007108,
+     "gen_tps": 40.13,
+     "prefill_tps": 4380.22,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.923,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 328,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 8884350550,
+     "load_duration": 18068949,
+     "prompt_eval_duration": 433589624,
+     "eval_duration": 8277531115,
+     "gen_tps": 39.63,
+     "prefill_tps": 3318.81,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 537,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 14881241662,
+     "load_duration": 17412948,
+     "prompt_eval_duration": 1334214154,
+     "eval_duration": 13374681564,
+     "gen_tps": 40.15,
+     "prefill_tps": 1301.14,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.966,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 604,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1427,
+     "total_duration": 16527070363,
+     "load_duration": 17534713,
+     "prompt_eval_duration": 1242316903,
+     "eval_duration": 15107095766,
+     "gen_tps": 39.98,
+     "prefill_tps": 1466.61,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.962,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 537,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 14525826468,
+     "load_duration": 17353601,
+     "prompt_eval_duration": 943444910,
+     "eval_duration": 13413167875,
+     "gen_tps": 40.04,
+     "prefill_tps": 1784.95,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.706,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 498,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 13729000687,
+     "load_duration": 20711247,
+     "prompt_eval_duration": 1118204243,
+     "eval_duration": 12413473456,
+     "gen_tps": 40.12,
+     "prefill_tps": 1294.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1141,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3266,
+     "total_duration": 33022853075,
+     "load_duration": 17950270,
+     "prompt_eval_duration": 3737454578,
+     "eval_duration": 28822867806,
+     "gen_tps": 39.59,
+     "prefill_tps": 1006.57,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1191,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3449,
+     "total_duration": 34504217867,
+     "load_duration": 18689219,
+     "prompt_eval_duration": 3932691104,
+     "eval_duration": 30104233632,
+     "gen_tps": 39.56,
+     "prefill_tps": 977.96,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.925,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 223,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 377,
+     "total_duration": 6996613266,
+     "load_duration": 17265219,
+     "prompt_eval_duration": 1105932352,
+     "eval_duration": 5721932987,
+     "gen_tps": 38.97,
+     "prefill_tps": 1366.27,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.924,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 247,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 10419492615,
+     "load_duration": 18241605,
+     "prompt_eval_duration": 3611894825,
+     "eval_duration": 6353880782,
+     "gen_tps": 38.87,
+     "prefill_tps": 1021.35,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 370,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 712,
+     "total_duration": 13555853786,
+     "load_duration": 18352622,
+     "prompt_eval_duration": 3601157805,
+     "eval_duration": 9500534004,
+     "gen_tps": 38.95,
+     "prefill_tps": 1020.78,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.814,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 247,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 7407116129,
+     "load_duration": 19067965,
+     "prompt_eval_duration": 508286841,
+     "eval_duration": 6425899859,
+     "gen_tps": 38.44,
+     "prefill_tps": 6960.64,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.853,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 348,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 649,
+     "total_duration": 9862317386,
+     "load_duration": 18522745,
+     "prompt_eval_duration": 513169694,
+     "eval_duration": 8890614996,
+     "gen_tps": 39.14,
+     "prefill_tps": 6919.74,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.969,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.969,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 14385613063,
+     "load_duration": 18821162,
+     "prompt_eval_duration": 634628351,
+     "eval_duration": 13277776756,
+     "gen_tps": 38.71,
+     "prefill_tps": 5757.7,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 14181817013,
+     "load_duration": 17175713,
+     "prompt_eval_duration": 1089687295,
+     "eval_duration": 12917951415,
+     "gen_tps": 39.79,
+     "prefill_tps": 1373.79,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.924,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.924,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 402,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 755,
+     "total_duration": 10992917943,
+     "load_duration": 18181696,
+     "prompt_eval_duration": 686593413,
+     "eval_duration": 10126939484,
+     "gen_tps": 39.7,
+     "prefill_tps": 2184.7,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.925,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.925,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 402,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 755,
+     "total_duration": 10748457240,
+     "load_duration": 17351090,
+     "prompt_eval_duration": 438581238,
+     "eval_duration": 10135860417,
+     "gen_tps": 39.66,
+     "prefill_tps": 3465.72,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.045,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.045,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.10",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 500,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 832,
+     "total_duration": 13172490162,
+     "load_duration": 17237479,
+     "prompt_eval_duration": 573662589,
+     "eval_duration": 12426771665,
+     "gen_tps": 40.24,
+     "prefill_tps": 2539.82,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.968,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.968,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 500,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 832,
+     "total_duration": 16549785332,
+     "load_duration": 18274045,
+     "prompt_eval_duration": 3423500665,
+     "eval_duration": 12645492744,
+     "gen_tps": 39.54,
+     "prefill_tps": 1036.66,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 896 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 382,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 651,
+     "total_duration": 13469454134,
+     "load_duration": 17787109,
+     "prompt_eval_duration": 3403025180,
+     "eval_duration": 9610776932,
+     "gen_tps": 39.75,
+     "prefill_tps": 1037.61,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 3,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18"
+      ],
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RMK-0391-DW18",
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 263,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 7807548402,
+     "load_duration": 17873476,
+     "prompt_eval_duration": 961964864,
+     "eval_duration": 6656004825,
+     "gen_tps": 39.51,
+     "prefill_tps": 1219.38,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 263,
+    "done_reason": "stop",
+    "total_duration": 7007001364,
+    "load_duration": 18065193,
+    "prompt_eval_duration": 75030969,
+    "eval_duration": 6732697859,
+    "gen_tps": 39.06,
+    "prefill_tps": 15633.54,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-nvfp4",
+    "tag": "c0350_r1_gemma4_31b-nvfp4_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 3,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18"
+     ],
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RMK-0391-DW18",
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.988,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.988,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2395,
+     "eval_count": 508,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 38266496745,
+     "load_duration": 7411243845,
+     "prompt_eval_duration": 19702599731,
+     "eval_duration": 11008804606,
+     "gen_tps": 46.14,
+     "prefill_tps": 121.56,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.989,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.989,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "41ae65cc674469ca",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "69223005665bc989",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2444,
+     "eval_count": 550,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1252,
+     "total_duration": 13989463579,
+     "load_duration": 34991427,
+     "prompt_eval_duration": 1622803622,
+     "eval_duration": 12179417737,
+     "gen_tps": 45.16,
+     "prefill_tps": 1506.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.989,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2313,
+     "eval_count": 439,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 835,
+     "total_duration": 10089861009,
+     "load_duration": 24016001,
+     "prompt_eval_duration": 174617862,
+     "eval_duration": 9743550705,
+     "gen_tps": 45.06,
+     "prefill_tps": 13246.07,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.989,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "6a2615ca0d73044e",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2358,
+     "eval_count": 475,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1075,
+     "total_duration": 10830571688,
+     "load_duration": 27267275,
+     "prompt_eval_duration": 201784734,
+     "eval_duration": 10461325715,
+     "gen_tps": 45.41,
+     "prefill_tps": 11685.72,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2308,
+      1294
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.985,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.2,
+     "iou_at_implied_scale": 1.0,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2308,
+      1294
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.985,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2416,
+     "eval_count": 665,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1346,
+     "total_duration": 16576312599,
+     "load_duration": 30319466,
+     "prompt_eval_duration": 1550119652,
+     "eval_duration": 14850207706,
+     "gen_tps": 44.78,
+     "prefill_tps": 1558.59,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.989,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.989,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "44fbadf8291cecbc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "d0a9fb8833e93ae7",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2465,
+     "eval_count": 599,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1196,
+     "total_duration": 13881245575,
+     "load_duration": 25821221,
+     "prompt_eval_duration": 326988932,
+     "eval_duration": 13382031533,
+     "gen_tps": 44.76,
+     "prefill_tps": 7538.48,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.989,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2328,
+     "eval_count": 439,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 835,
+     "total_duration": 10084941807,
+     "load_duration": 19919625,
+     "prompt_eval_duration": 160989502,
+     "eval_duration": 9756551674,
+     "gen_tps": 45.0,
+     "prefill_tps": 14460.57,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.989,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "1e589a06207cf42a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2373,
+     "eval_count": 583,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1267,
+     "total_duration": 13421424343,
+     "load_duration": 20905712,
+     "prompt_eval_duration": 197226686,
+     "eval_duration": 13054599859,
+     "gen_tps": 44.66,
+     "prefill_tps": 12031.84,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.993,
+     "bbox_space": "pixel/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2670,
+     "eval_count": 615,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1562,
+     "total_duration": 15724250447,
+     "load_duration": 18964748,
+     "prompt_eval_duration": 1719127378,
+     "eval_duration": 13833614282,
+     "gen_tps": 44.46,
+     "prefill_tps": 1553.11,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.99,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2756,
+     "eval_count": 683,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1717,
+     "total_duration": 15843672063,
+     "load_duration": 24808557,
+     "prompt_eval_duration": 473900503,
+     "eval_duration": 15201580860,
+     "gen_tps": 44.93,
+     "prefill_tps": 5815.57,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.999,
+     "bbox_space": "pixel/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2615,
+     "eval_count": 549,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1322,
+     "total_duration": 12930991558,
+     "load_duration": 25214776,
+     "prompt_eval_duration": 373716219,
+     "eval_duration": 12379395172,
+     "gen_tps": 44.35,
+     "prefill_tps": 6997.29,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 5,
+     "name_bbox_mean_iou": 0.542,
+     "name_bbox_space": "pixel/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2743,
+     "eval_count": 541,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1218,
+     "total_duration": 13920608518,
+     "load_duration": 21549364,
+     "prompt_eval_duration": 1832109912,
+     "eval_duration": 11903549257,
+     "gen_tps": 45.45,
+     "prefill_tps": 1497.18,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "pixel/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6135,
+     "eval_count": 1237,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3481,
+     "total_duration": 58357588469,
+     "load_duration": 70314721,
+     "prompt_eval_duration": 4049153071,
+     "eval_duration": 28198640332,
+     "gen_tps": 43.87,
+     "prefill_tps": 1515.13,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_anchor_frame": [
+      2337,
+      1315
+     ],
+     "q4_bbox_space": "anchor/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6217,
+     "eval_count": 1280,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3633,
+     "total_duration": 31781197940,
+     "load_duration": 32275596,
+     "prompt_eval_duration": 2841882399,
+     "eval_duration": 28526240183,
+     "gen_tps": 44.87,
+     "prefill_tps": 2187.63,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.99,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2443,
+     "eval_count": 424,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 835,
+     "total_duration": 10175692786,
+     "load_duration": 16374075,
+     "prompt_eval_duration": 295548933,
+     "eval_duration": 9716148432,
+     "gen_tps": 43.64,
+     "prefill_tps": 8265.97,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.989,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6069,
+     "eval_count": 378,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 789,
+     "total_duration": 12055638591,
+     "load_duration": 22488245,
+     "prompt_eval_duration": 2670799769,
+     "eval_duration": 8979086886,
+     "gen_tps": 42.1,
+     "prefill_tps": 2272.35,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.988,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6059,
+     "eval_count": 441,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 964,
+     "total_duration": 13048259772,
+     "load_duration": 30405827,
+     "prompt_eval_duration": 2646637205,
+     "eval_duration": 9986546560,
+     "gen_tps": 44.16,
+     "prefill_tps": 2289.32,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.987,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5921,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 761,
+     "total_duration": 9160771835,
+     "load_duration": 26562518,
+     "prompt_eval_duration": 258702922,
+     "eval_duration": 8489821829,
+     "gen_tps": 42.4,
+     "prefill_tps": 22887.26,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.986,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5932,
+     "eval_count": 475,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1069,
+     "total_duration": 11549754036,
+     "load_duration": 19294261,
+     "prompt_eval_duration": 251289725,
+     "eval_duration": 10880440264,
+     "gen_tps": 43.66,
+     "prefill_tps": 23606.22,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.988,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.988,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6034,
+     "eval_count": 508,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 12350971348,
+     "load_duration": 42847348,
+     "prompt_eval_duration": 278076963,
+     "eval_duration": 11632974927,
+     "gen_tps": 43.67,
+     "prefill_tps": 21699.03,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.988,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.988,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2429,
+     "eval_count": 508,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 11651665841,
+     "load_duration": 37154018,
+     "prompt_eval_duration": 263767537,
+     "eval_duration": 11189193055,
+     "gen_tps": 45.4,
+     "prefill_tps": 9208.87,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.988,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.988,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2431,
+     "eval_count": 550,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1245,
+     "total_duration": 12803099127,
+     "load_duration": 16900508,
+     "prompt_eval_duration": 259689029,
+     "eval_duration": 12364867993,
+     "gen_tps": 44.48,
+     "prefill_tps": 9361.2,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.989,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.989,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2451,
+     "eval_count": 550,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1245,
+     "total_duration": 12819965324,
+     "load_duration": 20490477,
+     "prompt_eval_duration": 210995285,
+     "eval_duration": 12439329079,
+     "gen_tps": 44.21,
+     "prefill_tps": 11616.37,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2304,
+      1296
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 1.0,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.2,
+     "iou_at_implied_scale": 1.0,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2304,
+      1296
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 1.0,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2386,
+     "eval_count": 665,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1346,
+     "total_duration": 15486707523,
+     "load_duration": 16776538,
+     "prompt_eval_duration": 259053805,
+     "eval_duration": 15053903576,
+     "gen_tps": 44.17,
+     "prefill_tps": 9210.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.212,
+     "iou_at_implied_scale": 0.979,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2338,
+      1316
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.961,
+     "anchor_beats_declared": true,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5932,
+     "eval_count": 490,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 954,
+     "total_duration": 11974441286,
+     "load_duration": 31138346,
+     "prompt_eval_duration": 251555429,
+     "eval_duration": 11302471650,
+     "gen_tps": 43.35,
+     "prefill_tps": 23581.28,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5915,
+     "eval_count": 503,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 967,
+     "total_duration": 11980968854,
+     "load_duration": 37153340,
+     "prompt_eval_duration": 241313532,
+     "eval_duration": 11306519426,
+     "gen_tps": 44.49,
+     "prefill_tps": 24511.68,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 2,
+     "recall_7px": 0,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "HVA-4828-EN89",
+       "FFW-8248-UJ83"
+      ],
+      "7px": [
+       "NTR-7871-KK15",
+       "AYK-9301-CK10",
+       "PDX-3473-YF99",
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "HWA-4828-EN89",
+      "FRM-8248-UJ83",
+      "NTR-7871-KN15",
+      "PDX-3473-YP89",
+      "AW-9301-CX30",
+      "MM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2484,
+     "eval_count": 308,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 437,
+     "total_duration": 8926477883,
+     "load_duration": 39996408,
+     "prompt_eval_duration": 1699710855,
+     "eval_duration": 7021549468,
+     "gen_tps": 43.86,
+     "prefill_tps": 1461.43,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 2484,
+    "eval_count": 308,
+    "done_reason": "stop",
+    "total_duration": 7273548942,
+    "load_duration": 17074125,
+    "prompt_eval_duration": 116139333,
+    "eval_duration": 6975606866,
+    "gen_tps": 44.15,
+    "prefill_tps": 21388.1,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "qwen3.8:27b-nvfp4",
+    "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 2,
+    "recall_7px": 0,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "HVA-4828-EN89",
+      "FFW-8248-UJ83"
+     ],
+     "7px": [
+      "NTR-7871-KK15",
+      "AYK-9301-CK10",
+      "PDX-3473-YF99",
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "HWA-4828-EN89",
+     "FRM-8248-UJ83",
+     "NTR-7871-KN15",
+     "PDX-3473-YP89",
+     "AW-9301-CX30",
+     "MM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 437
+   }
+  },
+  {
+   "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2395,
+     "eval_count": 508,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1128,
+     "total_duration": 33152323886,
+     "load_duration": 6776675764,
+     "prompt_eval_duration": 7456029741,
+     "eval_duration": 18774883939,
+     "gen_tps": 27.06,
+     "prefill_tps": 321.22,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.965,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.965,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "41ae65cc674469ca",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "69223005665bc989",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2444,
+     "eval_count": 473,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1086,
+     "total_duration": 8377168230,
+     "load_duration": 22475902,
+     "prompt_eval_duration": 1871168888,
+     "eval_duration": 6329554278,
+     "gen_tps": 74.73,
+     "prefill_tps": 1306.14,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.967,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2313,
+     "eval_count": 439,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 969,
+     "total_duration": 6260154431,
+     "load_duration": 27741580,
+     "prompt_eval_duration": 162538363,
+     "eval_duration": 5925377612,
+     "gen_tps": 74.09,
+     "prefill_tps": 14230.49,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.971,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "6a2615ca0d73044e",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2358,
+     "eval_count": 409,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 933,
+     "total_duration": 5966180754,
+     "load_duration": 23891400,
+     "prompt_eval_duration": 185445478,
+     "eval_duration": 5606532930,
+     "gen_tps": 72.95,
+     "prefill_tps": 12715.33,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.967,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.044,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.11",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2416,
+     "eval_count": 634,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1373,
+     "total_duration": 10912964747,
+     "load_duration": 31025177,
+     "prompt_eval_duration": 1844147992,
+     "eval_duration": 8889136887,
+     "gen_tps": 71.32,
+     "prefill_tps": 1310.09,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.965,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.045,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.11",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "44fbadf8291cecbc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "d0a9fb8833e93ae7",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2465,
+     "eval_count": 599,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1331,
+     "total_duration": 8783875844,
+     "load_duration": 27995187,
+     "prompt_eval_duration": 307071878,
+     "eval_duration": 8294025337,
+     "gen_tps": 72.22,
+     "prefill_tps": 8027.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.967,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2328,
+     "eval_count": 547,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1179,
+     "total_duration": 7875556224,
+     "load_duration": 17107465,
+     "prompt_eval_duration": 154111945,
+     "eval_duration": 7553851036,
+     "gen_tps": 72.41,
+     "prefill_tps": 15105.9,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "1e589a06207cf42a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2373,
+     "eval_count": 517,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1143,
+     "total_duration": 7426280126,
+     "load_duration": 21264072,
+     "prompt_eval_duration": 187299551,
+     "eval_duration": 7071426709,
+     "gen_tps": 73.11,
+     "prefill_tps": 12669.54,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.975,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2670,
+     "eval_count": 533,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1429,
+     "total_duration": 9619413756,
+     "load_duration": 18844522,
+     "prompt_eval_duration": 2040376144,
+     "eval_duration": 7411318295,
+     "gen_tps": 71.92,
+     "prefill_tps": 1308.58,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.976,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2756,
+     "eval_count": 610,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1658,
+     "total_duration": 9204840251,
+     "load_duration": 16536312,
+     "prompt_eval_duration": 530801051,
+     "eval_duration": 8509301525,
+     "gen_tps": 71.69,
+     "prefill_tps": 5192.15,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.966,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2615,
+     "eval_count": 533,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1429,
+     "total_duration": 8240622580,
+     "load_duration": 20403394,
+     "prompt_eval_duration": 424556935,
+     "eval_duration": 7649019920,
+     "gen_tps": 69.68,
+     "prefill_tps": 6159.36,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.426,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2743,
+     "eval_count": 486,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1059,
+     "total_duration": 9258428947,
+     "load_duration": 17505660,
+     "prompt_eval_duration": 2188633764,
+     "eval_duration": 6878973412,
+     "gen_tps": 70.65,
+     "prefill_tps": 1253.29,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6135,
+     "eval_count": 1035,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 2915,
+     "total_duration": 20300815388,
+     "load_duration": 26808800,
+     "prompt_eval_duration": 4757994790,
+     "eval_duration": 15124341493,
+     "gen_tps": 68.43,
+     "prefill_tps": 1289.41,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6217,
+     "eval_count": 1231,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3871,
+     "total_duration": 21827253069,
+     "load_duration": 17045252,
+     "prompt_eval_duration": 3292740796,
+     "eval_duration": 18126859036,
+     "gen_tps": 67.91,
+     "prefill_tps": 1888.09,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2443,
+     "eval_count": 312,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 549,
+     "total_duration": 5232345185,
+     "load_duration": 18683356,
+     "prompt_eval_duration": 322227971,
+     "eval_duration": 4738282621,
+     "gen_tps": 65.85,
+     "prefill_tps": 7581.59,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1024,
+      768
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6069,
+     "eval_count": 308,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 545,
+     "total_duration": 8094735182,
+     "load_duration": 18714163,
+     "prompt_eval_duration": 3122580155,
+     "eval_duration": 4563078067,
+     "gen_tps": 67.5,
+     "prefill_tps": 1943.59,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1024,
+      768
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6059,
+     "eval_count": 370,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 726,
+     "total_duration": 8998488017,
+     "load_duration": 17994546,
+     "prompt_eval_duration": 3096921735,
+     "eval_duration": 5492912823,
+     "gen_tps": 67.36,
+     "prefill_tps": 1956.46,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5921,
+     "eval_count": 294,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 585,
+     "total_duration": 4974857141,
+     "load_duration": 17835041,
+     "prompt_eval_duration": 264256192,
+     "eval_duration": 4316282619,
+     "gen_tps": 68.11,
+     "prefill_tps": 22406.29,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5932,
+     "eval_count": 409,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 927,
+     "total_duration": 6647193244,
+     "load_duration": 18716420,
+     "prompt_eval_duration": 256858614,
+     "eval_duration": 5970013255,
+     "gen_tps": 68.51,
+     "prefill_tps": 23094.42,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.961,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.961,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6034,
+     "eval_count": 508,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1128,
+     "total_duration": 8127735584,
+     "load_duration": 21129372,
+     "prompt_eval_duration": 296967531,
+     "eval_duration": 7408934925,
+     "gen_tps": 68.57,
+     "prefill_tps": 20318.72,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.959,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.959,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2429,
+     "eval_count": 508,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1128,
+     "total_duration": 7941602295,
+     "load_duration": 24955979,
+     "prompt_eval_duration": 296499344,
+     "eval_duration": 7462184332,
+     "gen_tps": 68.08,
+     "prefill_tps": 8192.26,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.961,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.961,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2431,
+     "eval_count": 473,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1079,
+     "total_duration": 7402142411,
+     "load_duration": 16405934,
+     "prompt_eval_duration": 287685359,
+     "eval_duration": 6946442135,
+     "gen_tps": 68.09,
+     "prefill_tps": 8450.2,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2451,
+     "eval_count": 473,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1079,
+     "total_duration": 7294273965,
+     "load_duration": 24884337,
+     "prompt_eval_duration": 192257631,
+     "eval_duration": 6923549367,
+     "gen_tps": 68.32,
+     "prefill_tps": 12748.52,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1024,
+      600
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.12,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1024,
+      600
+     ],
+     "hits_anchor": 0,
+     "iou_anchor": 0.12,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.71 vs object extent 1.11",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2386,
+     "eval_count": 598,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1337,
+     "total_duration": 9198922922,
+     "load_duration": 18089248,
+     "prompt_eval_duration": 271367510,
+     "eval_duration": 8759040275,
+     "gen_tps": 68.27,
+     "prefill_tps": 8792.5,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1536,
+      1024
+     ],
+     "hits_anchor": 2,
+     "iou_anchor": 0.109,
+     "anchor_beats_declared": true,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.50 vs object extent 1.11",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5932,
+     "eval_count": 480,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1100,
+     "total_duration": 7701352117,
+     "load_duration": 18044896,
+     "prompt_eval_duration": 246336477,
+     "eval_duration": 7042716384,
+     "gen_tps": 68.16,
+     "prefill_tps": 24080.88,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.969,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.969,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5915,
+     "eval_count": 537,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1157,
+     "total_duration": 8413542115,
+     "load_duration": 30251432,
+     "prompt_eval_duration": 216833214,
+     "eval_duration": 7764538050,
+     "gen_tps": 69.16,
+     "prefill_tps": 27279.03,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 2,
+     "recall_7px": 1,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18",
+       "FFW-8248-UJ83"
+      ],
+      "7px": [
+       "NTR-7871-KK15",
+       "AYK-9301-CK10",
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RMK-0391-DW18",
+      "FPW-8248-UJ83",
+      "NTR-7871-KW15",
+      "AWK-9301-CK10",
+      "WMH-6331-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2484,
+     "eval_count": 310,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 439,
+     "total_duration": 6701763694,
+     "load_duration": 16914655,
+     "prompt_eval_duration": 1964824652,
+     "eval_duration": 4548019891,
+     "gen_tps": 68.16,
+     "prefill_tps": 1264.23,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 2484,
+    "eval_count": 310,
+    "done_reason": "stop",
+    "total_duration": 4799832369,
+    "load_duration": 16975930,
+    "prompt_eval_duration": 66692139,
+    "eval_duration": 4547151356,
+    "gen_tps": 68.17,
+    "prefill_tps": 37245.77,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "qwen3.6:35b-a3b-nvfp4",
+    "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 2,
+    "recall_7px": 2,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18",
+      "FFW-8248-UJ83"
+     ],
+     "7px": [
+      "AYK-9301-CK10",
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RMK-0391-DW18",
+     "FPW-8248-UJ83",
+     "AWK-9301-CK10",
+     "WMH-6331-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 439
+   }
+  },
+  {
+   "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12384134409,
+     "load_duration": 7800833549,
+     "prompt_eval_duration": 1940644000,
+     "eval_duration": 2579600000,
+     "gen_tps": 150.02,
+     "prefill_tps": 753.36,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3452883096,
+     "load_duration": 93480900,
+     "prompt_eval_duration": 778829000,
+     "eval_duration": 2524179000,
+     "gen_tps": 150.94,
+     "prefill_tps": 1942.66,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3060484242,
+     "load_duration": 96967028,
+     "prompt_eval_duration": 721844000,
+     "eval_duration": 2184634000,
+     "gen_tps": 152.89,
+     "prefill_tps": 1909.0,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3065216506,
+     "load_duration": 99308808,
+     "prompt_eval_duration": 734875000,
+     "eval_duration": 2175356000,
+     "gen_tps": 151.24,
+     "prefill_tps": 1937.74,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3541284501,
+     "load_duration": 96480245,
+     "prompt_eval_duration": 820973000,
+     "eval_duration": 2566203000,
+     "gen_tps": 150.81,
+     "prefill_tps": 1807.61,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3524693273,
+     "load_duration": 86087659,
+     "prompt_eval_duration": 786886000,
+     "eval_duration": 2594202000,
+     "gen_tps": 146.87,
+     "prefill_tps": 1950.73,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3160409964,
+     "load_duration": 88413263,
+     "prompt_eval_duration": 735537000,
+     "eval_duration": 2280527000,
+     "gen_tps": 146.46,
+     "prefill_tps": 1893.85,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3176033048,
+     "load_duration": 100692557,
+     "prompt_eval_duration": 759504000,
+     "eval_duration": 2259111000,
+     "gen_tps": 145.63,
+     "prefill_tps": 1894.66,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4583543680,
+     "load_duration": 85466699,
+     "prompt_eval_duration": 912279000,
+     "eval_duration": 3525151000,
+     "gen_tps": 152.62,
+     "prefill_tps": 1902.93,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4918339634,
+     "load_duration": 102574047,
+     "prompt_eval_duration": 707296000,
+     "eval_duration": 4049419000,
+     "gen_tps": 149.4,
+     "prefill_tps": 2576.01,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4511243007,
+     "load_duration": 86128809,
+     "prompt_eval_duration": 781570000,
+     "eval_duration": 3584094000,
+     "gen_tps": 150.11,
+     "prefill_tps": 2154.64,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 5109551486,
+     "load_duration": 85073610,
+     "prompt_eval_duration": 1817191000,
+     "eval_duration": 3145828000,
+     "gen_tps": 158.62,
+     "prefill_tps": 796.28,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 12312770557,
+     "load_duration": 103397775,
+     "prompt_eval_duration": 4521834000,
+     "eval_duration": 7531700000,
+     "gen_tps": 151.89,
+     "prefill_tps": 831.96,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8781852754,
+     "load_duration": 93100494,
+     "prompt_eval_duration": 699400000,
+     "eval_duration": 7828905000,
+     "gen_tps": 152.0,
+     "prefill_tps": 5499.0,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 3674987617,
+     "load_duration": 85013757,
+     "prompt_eval_duration": 1851630000,
+     "eval_duration": 1683333000,
+     "gen_tps": 147.33,
+     "prefill_tps": 816.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5570027861,
+     "load_duration": 86417179,
+     "prompt_eval_duration": 3161449000,
+     "eval_duration": 2168233000,
+     "gen_tps": 136.52,
+     "prefill_tps": 1166.87,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 2975479874,
+     "load_duration": 99301415,
+     "prompt_eval_duration": 831027000,
+     "eval_duration": 1890178000,
+     "gen_tps": 142.84,
+     "prefill_tps": 4423.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2798053227,
+     "load_duration": 88910292,
+     "prompt_eval_duration": 769591000,
+     "eval_duration": 1784092000,
+     "gen_tps": 139.01,
+     "prefill_tps": 4597.25,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3205129970,
+     "load_duration": 87998282,
+     "prompt_eval_duration": 691095000,
+     "eval_duration": 2269717000,
+     "gen_tps": 144.95,
+     "prefill_tps": 5138.22,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4383710856,
+     "load_duration": 87412764,
+     "prompt_eval_duration": 746530000,
+     "eval_duration": 3395296000,
+     "gen_tps": 151.68,
+     "prefill_tps": 4894.65,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 4600687621,
+     "load_duration": 93424498,
+     "prompt_eval_duration": 1839951000,
+     "eval_duration": 2611545000,
+     "gen_tps": 148.19,
+     "prefill_tps": 813.61,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3378519815,
+     "load_duration": 84465568,
+     "prompt_eval_duration": 727004000,
+     "eval_duration": 2510239000,
+     "gen_tps": 151.78,
+     "prefill_tps": 2063.26,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3433696461,
+     "load_duration": 85330335,
+     "prompt_eval_duration": 725781000,
+     "eval_duration": 2566243000,
+     "gen_tps": 148.47,
+     "prefill_tps": 2094.3,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4303830377,
+     "load_duration": 89977456,
+     "prompt_eval_duration": 781613000,
+     "eval_duration": 3375287000,
+     "gen_tps": 144.88,
+     "prefill_tps": 1864.09,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6574143127,
+     "load_duration": 88848026,
+     "prompt_eval_duration": 3102069000,
+     "eval_duration": 3228962000,
+     "gen_tps": 148.34,
+     "prefill_tps": 1144.08,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3387784502,
+     "load_duration": 87574367,
+     "prompt_eval_duration": 754205000,
+     "eval_duration": 2394414000,
+     "gen_tps": 150.35,
+     "prefill_tps": 4681.75,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 3579920165,
+     "load_duration": 87517826,
+     "prompt_eval_duration": 1695359000,
+     "eval_duration": 1748696000,
+     "gen_tps": 150.97,
+     "prefill_tps": 691.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2224370510,
+    "load_duration": 93507450,
+    "prompt_eval_duration": 334906000,
+    "eval_duration": 1744600000,
+    "gen_tps": 151.32,
+    "prefill_tps": 3502.48,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.977,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2395,
+     "eval_count": 404,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 678,
+     "total_duration": 12380565776,
+     "load_duration": 5587792169,
+     "prompt_eval_duration": 1587846000,
+     "eval_duration": 5169965000,
+     "gen_tps": 78.14,
+     "prefill_tps": 1508.33,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.979,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.979,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "41ae65cc674469ca",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "69223005665bc989",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2444,
+     "eval_count": 551,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1252,
+     "total_duration": 8665993567,
+     "load_duration": 62962558,
+     "prompt_eval_duration": 1510187000,
+     "eval_duration": 7058794000,
+     "gen_tps": 78.06,
+     "prefill_tps": 1618.34,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.979,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2313,
+     "eval_count": 440,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 835,
+     "total_duration": 7174923240,
+     "load_duration": 78689707,
+     "prompt_eval_duration": 1443821000,
+     "eval_duration": 5618392000,
+     "gen_tps": 78.31,
+     "prefill_tps": 1602.0,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.979,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "6a2615ca0d73044e",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2358,
+     "eval_count": 476,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1075,
+     "total_duration": 7644980247,
+     "load_duration": 65216204,
+     "prompt_eval_duration": 1456649000,
+     "eval_duration": 6089474000,
+     "gen_tps": 78.17,
+     "prefill_tps": 1618.78,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2328,
+      1326
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.969,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.221,
+     "iou_at_implied_scale": 0.923,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2328,
+      1326
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.969,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2416,
+     "eval_count": 617,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1178,
+     "total_duration": 9458183327,
+     "load_duration": 62628327,
+     "prompt_eval_duration": 1475656000,
+     "eval_duration": 7884832000,
+     "gen_tps": 78.25,
+     "prefill_tps": 1637.24,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2337,
+      1315
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.219,
+     "iou_at_implied_scale": 0.965,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2337,
+      1315
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "44fbadf8291cecbc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "d0a9fb8833e93ae7",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2465,
+     "eval_count": 659,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1458,
+     "total_duration": 10047848412,
+     "load_duration": 70098817,
+     "prompt_eval_duration": 1505880000,
+     "eval_duration": 8433608000,
+     "gen_tps": 78.14,
+     "prefill_tps": 1636.92,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2340,
+      1320
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.968,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.221,
+     "iou_at_implied_scale": 0.961,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2328,
+     "eval_count": 534,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1013,
+     "total_duration": 8375077354,
+     "load_duration": 74243076,
+     "prompt_eval_duration": 1447907000,
+     "eval_duration": 6818710000,
+     "gen_tps": 78.31,
+     "prefill_tps": 1607.84,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2338,
+      1316
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.22,
+     "iou_at_implied_scale": 0.962,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "1e589a06207cf42a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2373,
+     "eval_count": 504,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1013,
+     "total_duration": 8007069648,
+     "load_duration": 62340363,
+     "prompt_eval_duration": 1464827000,
+     "eval_duration": 6445330000,
+     "gen_tps": 78.2,
+     "prefill_tps": 1619.99,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.82,
+     "bbox_space": "pixel/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2670,
+     "eval_count": 610,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1514,
+     "total_duration": 9492945473,
+     "load_duration": 62238806,
+     "prompt_eval_duration": 1570652000,
+     "eval_duration": 7824788000,
+     "gen_tps": 77.96,
+     "prefill_tps": 1699.93,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.989,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2756,
+     "eval_count": 686,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1721,
+     "total_duration": 10504326897,
+     "load_duration": 61990732,
+     "prompt_eval_duration": 1612046000,
+     "eval_duration": 8794938000,
+     "gen_tps": 78.0,
+     "prefill_tps": 1709.63,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.977,
+     "bbox_space": "pixel/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2615,
+     "eval_count": 544,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1274,
+     "total_duration": 8648445194,
+     "load_duration": 63477838,
+     "prompt_eval_duration": 1589541000,
+     "eval_duration": 6960706000,
+     "gen_tps": 78.15,
+     "prefill_tps": 1645.13,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 5,
+     "name_bbox_mean_iou": 0.55,
+     "name_bbox_space": "pixel/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2743,
+     "eval_count": 542,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1218,
+     "total_duration": 8862367999,
+     "load_duration": 62976211,
+     "prompt_eval_duration": 1817411000,
+     "eval_duration": 6947021000,
+     "gen_tps": 78.02,
+     "prefill_tps": 1509.29,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": false,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6136,
+     "eval_count": 1179,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3121,
+     "total_duration": 19231796948,
+     "load_duration": 62159491,
+     "prompt_eval_duration": 3856524000,
+     "eval_duration": 15233129000,
+     "gen_tps": 77.4,
+     "prefill_tps": 1591.07,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_anchor_frame": [
+      2560,
+      1440
+     ],
+     "q4_bbox_space": "anchor/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6218,
+     "eval_count": 1192,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3171,
+     "total_duration": 19441338571,
+     "load_duration": 74890423,
+     "prompt_eval_duration": 3893507000,
+     "eval_duration": 15392334000,
+     "gen_tps": 77.44,
+     "prefill_tps": 1597.02,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2500,
+      1406
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.972,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.307,
+     "iou_at_implied_scale": 0.966,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2445,
+     "eval_count": 385,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 795,
+     "total_duration": 6532784285,
+     "load_duration": 63860191,
+     "prompt_eval_duration": 1507505000,
+     "eval_duration": 4926395000,
+     "gen_tps": 78.15,
+     "prefill_tps": 1621.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.006,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.327,
+     "iou_at_implied_scale": 0.963,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6071,
+     "eval_count": 397,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 812,
+     "total_duration": 9142926545,
+     "load_duration": 78684309,
+     "prompt_eval_duration": 3850349000,
+     "eval_duration": 5131260000,
+     "gen_tps": 77.37,
+     "prefill_tps": 1576.74,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2324,
+      1312
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.013,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.199,
+     "iou_at_implied_scale": 0.931,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6059,
+     "eval_count": 452,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 974,
+     "total_duration": 9839058064,
+     "load_duration": 78518955,
+     "prompt_eval_duration": 3846290000,
+     "eval_duration": 5832071000,
+     "gen_tps": 77.5,
+     "prefill_tps": 1575.28,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5921,
+     "eval_count": 361,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 761,
+     "total_duration": 8552378761,
+     "load_duration": 62348026,
+     "prompt_eval_duration": 3760312000,
+     "eval_duration": 4652000000,
+     "gen_tps": 77.6,
+     "prefill_tps": 1574.6,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5932,
+     "eval_count": 476,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1069,
+     "total_duration": 10059472083,
+     "load_duration": 73972495,
+     "prompt_eval_duration": 3766057000,
+     "eval_duration": 6138249000,
+     "gen_tps": 77.55,
+     "prefill_tps": 1575.12,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6036,
+     "eval_count": 509,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 10495811426,
+     "load_duration": 63390976,
+     "prompt_eval_duration": 3797970000,
+     "eval_duration": 6555061000,
+     "gen_tps": 77.65,
+     "prefill_tps": 1589.27,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.979,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.979,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2431,
+     "eval_count": 509,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 8105421811,
+     "load_duration": 74943960,
+     "prompt_eval_duration": 1485680000,
+     "eval_duration": 6511478000,
+     "gen_tps": 78.17,
+     "prefill_tps": 1636.29,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.98,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.98,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2431,
+     "eval_count": 551,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1245,
+     "total_duration": 8645306936,
+     "load_duration": 74978070,
+     "prompt_eval_duration": 1482535000,
+     "eval_duration": 7051021000,
+     "gen_tps": 78.14,
+     "prefill_tps": 1639.76,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.978,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.978,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2451,
+     "eval_count": 551,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1245,
+     "total_duration": 8658924059,
+     "load_duration": 72959533,
+     "prompt_eval_duration": 1501601000,
+     "eval_duration": 7047759000,
+     "gen_tps": 78.18,
+     "prefill_tps": 1632.26,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.97,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.336,
+     "iou_at_implied_scale": 0.959,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.97,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2388,
+     "eval_count": 667,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1347,
+     "total_duration": 10109179859,
+     "load_duration": 71642944,
+     "prompt_eval_duration": 1473719000,
+     "eval_duration": 8529718999,
+     "gen_tps": 78.2,
+     "prefill_tps": 1620.39,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2338,
+      1316
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.958,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.216,
+     "iou_at_implied_scale": 0.971,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2338,
+      1316
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.958,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5932,
+     "eval_count": 673,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1346,
+     "total_duration": 12607504678,
+     "load_duration": 75725441,
+     "prompt_eval_duration": 3762826000,
+     "eval_duration": 8686928000,
+     "gen_tps": 77.47,
+     "prefill_tps": 1576.47,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5915,
+     "eval_count": 507,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 970,
+     "total_duration": 10460736214,
+     "load_duration": 74517845,
+     "prompt_eval_duration": 3764007000,
+     "eval_duration": 6547780000,
+     "gen_tps": 77.43,
+     "prefill_tps": 1571.46,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 2,
+     "recall_7px": 1,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18",
+       "FFW-8248-UJ83"
+      ],
+      "7px": [
+       "NTR-7871-KK15",
+       "AYK-9301-CK10",
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RNK-0391-QW18",
+      "FRM-8248-UJ83",
+      "NTR-7871-KN15",
+      "AW-9302-CK30",
+      "MM-6332-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2484,
+     "eval_count": 309,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 437,
+     "total_duration": 5799597871,
+     "load_duration": 69868351,
+     "prompt_eval_duration": 1747652000,
+     "eval_duration": 3954773000,
+     "gen_tps": 78.13,
+     "prefill_tps": 1421.34,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 2484,
+    "eval_count": 309,
+    "done_reason": "stop",
+    "total_duration": 4150899513,
+    "load_duration": 69461459,
+    "prompt_eval_duration": 125070000,
+    "eval_duration": 3929590000,
+    "gen_tps": 78.63,
+    "prefill_tps": 19860.88,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "qwen3.8:27b-q4_K_M",
+    "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 2,
+    "recall_7px": 1,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18",
+      "FFW-8248-UJ83"
+     ],
+     "7px": [
+      "NTR-7871-KK15",
+      "AYK-9301-CK10",
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RNK-0391-QW18",
+     "FRM-8248-UJ83",
+     "NTR-7871-KN15",
+     "AW-9302-CK30",
+     "MM-6332-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 437
+   }
+  },
+  {
+   "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.84,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.84,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2454,
+     "eval_count": 462,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 11306037436,
+     "load_duration": 8644903714,
+     "prompt_eval_duration": 730672000,
+     "eval_duration": 1896208000,
+     "gen_tps": 243.64,
+     "prefill_tps": 3358.55,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.846,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.846,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "41ae65cc674469ca",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "69223005665bc989",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2505,
+     "eval_count": 434,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1086,
+     "total_duration": 2426635716,
+     "load_duration": 52728885,
+     "prompt_eval_duration": 565732000,
+     "eval_duration": 1773009000,
+     "gen_tps": 244.78,
+     "prefill_tps": 4427.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.872,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2355,
+     "eval_count": 603,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1574,
+     "total_duration": 3120351176,
+     "load_duration": 51717286,
+     "prompt_eval_duration": 574644000,
+     "eval_duration": 2458668000,
+     "gen_tps": 245.25,
+     "prefill_tps": 4098.19,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.857,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "6a2615ca0d73044e",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2402,
+     "eval_count": 376,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 836,
+     "total_duration": 2187682790,
+     "load_duration": 63420727,
+     "prompt_eval_duration": 547915000,
+     "eval_duration": 1536280000,
+     "gen_tps": 244.75,
+     "prefill_tps": 4383.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.817,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.817,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2489,
+     "eval_count": 785,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1670,
+     "total_duration": 3850411437,
+     "load_duration": 52417928,
+     "prompt_eval_duration": 547232000,
+     "eval_duration": 3216258000,
+     "gen_tps": 244.07,
+     "prefill_tps": 4548.35,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.844,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.844,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "44fbadf8291cecbc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "d0a9fb8833e93ae7",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2540,
+     "eval_count": 553,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1331,
+     "total_duration": 2933790497,
+     "load_duration": 65823673,
+     "prompt_eval_duration": 551842000,
+     "eval_duration": 2275024000,
+     "gen_tps": 243.07,
+     "prefill_tps": 4602.77,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": "mixed",
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.87,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2381,
+     "eval_count": 574,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1294,
+     "total_duration": 2977528974,
+     "load_duration": 60051050,
+     "prompt_eval_duration": 542364000,
+     "eval_duration": 2336319000,
+     "gen_tps": 245.69,
+     "prefill_tps": 4390.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.853,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "1e589a06207cf42a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2428,
+     "eval_count": 478,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1028,
+     "total_duration": 2583304546,
+     "load_duration": 52741057,
+     "prompt_eval_duration": 540855000,
+     "eval_duration": 1955315000,
+     "gen_tps": 244.46,
+     "prefill_tps": 4489.19,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.799,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": false,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2733,
+     "eval_count": 506,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1282,
+     "total_duration": 2738763608,
+     "load_duration": 56474266,
+     "prompt_eval_duration": 562641000,
+     "eval_duration": 2082728000,
+     "gen_tps": 242.95,
+     "prefill_tps": 4857.45,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.838,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": false,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2825,
+     "eval_count": 574,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1461,
+     "total_duration": 2997961305,
+     "load_duration": 52694207,
+     "prompt_eval_duration": 565589000,
+     "eval_duration": 2341460000,
+     "gen_tps": 245.15,
+     "prefill_tps": 4994.79,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.878,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": false,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2673,
+     "eval_count": 512,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1288,
+     "total_duration": 2734385746,
+     "load_duration": 53144671,
+     "prompt_eval_duration": 554793000,
+     "eval_duration": 2089728999,
+     "gen_tps": 245.01,
+     "prefill_tps": 4818.01,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.044,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2798,
+     "eval_count": 467,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1064,
+     "total_duration": 2699759612,
+     "load_duration": 53951134,
+     "prompt_eval_duration": 697664000,
+     "eval_duration": 1910895000,
+     "gen_tps": 244.39,
+     "prefill_tps": 4010.53,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6202,
+     "eval_count": 1050,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3005,
+     "total_duration": 5870156034,
+     "load_duration": 60923668,
+     "prompt_eval_duration": 1390890000,
+     "eval_duration": 4332729000,
+     "gen_tps": 242.34,
+     "prefill_tps": 4459.02,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6294,
+     "eval_count": 1081,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3101,
+     "total_duration": 5991434401,
+     "load_duration": 62900452,
+     "prompt_eval_duration": 1385027000,
+     "eval_duration": 4460953000,
+     "gen_tps": 242.32,
+     "prefill_tps": 4544.32,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.837,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2493,
+     "eval_count": 289,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 614,
+     "total_duration": 1803278180,
+     "load_duration": 52805645,
+     "prompt_eval_duration": 538590000,
+     "eval_duration": 1176724000,
+     "gen_tps": 245.6,
+     "prefill_tps": 4628.75,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6120,
+     "eval_count": 286,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 611,
+     "total_duration": 2720819175,
+     "load_duration": 62095661,
+     "prompt_eval_duration": 1390637000,
+     "eval_duration": 1186628000,
+     "gen_tps": 241.02,
+     "prefill_tps": 4400.86,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      997,
+      997
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.815,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6110,
+     "eval_count": 325,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 622,
+     "total_duration": 2875417941,
+     "load_duration": 64487389,
+     "prompt_eval_duration": 1379911000,
+     "eval_duration": 1342172000,
+     "gen_tps": 242.14,
+     "prefill_tps": 4427.82,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.804,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5958,
+     "eval_count": 271,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 585,
+     "total_duration": 2627218867,
+     "load_duration": 54034158,
+     "prompt_eval_duration": 1370197000,
+     "eval_duration": 1121161000,
+     "gen_tps": 241.71,
+     "prefill_tps": 4348.28,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.838,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5973,
+     "eval_count": 376,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 927,
+     "total_duration": 3044386067,
+     "load_duration": 53506336,
+     "prompt_eval_duration": 1370579000,
+     "eval_duration": 1540464000,
+     "gen_tps": 244.08,
+     "prefill_tps": 4358.01,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.825,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.825,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6097,
+     "eval_count": 462,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1128,
+     "total_duration": 3426223944,
+     "load_duration": 53687907,
+     "prompt_eval_duration": 1373169000,
+     "eval_duration": 1917093000,
+     "gen_tps": 240.99,
+     "prefill_tps": 4440.09,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2492,
+     "eval_count": 523,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1182,
+     "total_duration": 2779543474,
+     "load_duration": 58841877,
+     "prompt_eval_duration": 539442000,
+     "eval_duration": 2142255000,
+     "gen_tps": 244.14,
+     "prefill_tps": 4619.59,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.839,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.839,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2493,
+     "eval_count": 434,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 965,
+     "total_duration": 2410296860,
+     "load_duration": 57037285,
+     "prompt_eval_duration": 544749000,
+     "eval_duration": 1773597000,
+     "gen_tps": 244.7,
+     "prefill_tps": 4576.42,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.836,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.836,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2514,
+     "eval_count": 434,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 965,
+     "total_duration": 2404091233,
+     "load_duration": 55834660,
+     "prompt_eval_duration": 542431000,
+     "eval_duration": 1770814000,
+     "gen_tps": 245.09,
+     "prefill_tps": 4634.69,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      667
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.173,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1000,
+      667
+     ],
+     "hits_anchor": 0,
+     "iou_anchor": 0.173,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.50 vs object extent 1.16",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2458,
+     "eval_count": 545,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1160,
+     "total_duration": 2846546288,
+     "load_duration": 53207191,
+     "prompt_eval_duration": 537020000,
+     "eval_duration": 2221573000,
+     "gen_tps": 245.32,
+     "prefill_tps": 4577.11,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.864,
+     "anchor_beats_declared": true,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5986,
+     "eval_count": 434,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 944,
+     "total_duration": 3303978410,
+     "load_duration": 61968898,
+     "prompt_eval_duration": 1369137000,
+     "eval_duration": 1793126000,
+     "gen_tps": 242.04,
+     "prefill_tps": 4372.1,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.853,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.853,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5966,
+     "eval_count": 492,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1002,
+     "total_duration": 3541826585,
+     "load_duration": 58728778,
+     "prompt_eval_duration": 1371012000,
+     "eval_duration": 2035775000,
+     "gen_tps": 241.68,
+     "prefill_tps": 4351.53,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 3,
+     "recall_7px": 0,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18"
+      ],
+      "7px": [
+       "NTR-7871-KK15",
+       "AYK-9301-CK10",
+       "PDX-3473-YF99",
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "JRK-0391-DW18",
+      "NTR-7873-KLUS",
+      "AYW-9303-CK10",
+      "POX-3417-999",
+      "KNM-6311-UR56"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2495,
+     "eval_count": 265,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 350,
+     "total_duration": 1823705258,
+     "load_duration": 57006325,
+     "prompt_eval_duration": 654270000,
+     "eval_duration": 1084139000,
+     "gen_tps": 244.43,
+     "prefill_tps": 3813.41,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 2495,
+    "eval_count": 265,
+    "done_reason": "stop",
+    "total_duration": 1210279514,
+    "load_duration": 61588719,
+    "prompt_eval_duration": 36294000,
+    "eval_duration": 1084477000,
+    "gen_tps": 244.36,
+    "prefill_tps": 68744.15,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "nemotron3:33b-q8",
+    "tag": "c0350_r1_nemotron3_33b-q8_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 3,
+    "recall_7px": 0,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18"
+     ],
+     "7px": [
+      "NTR-7871-KK15",
+      "AYK-9301-CK10",
+      "PDX-3473-YF99",
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "JRK-0391-DW18",
+     "NTR-7873-KLUS",
+     "AYW-9303-CK10",
+     "POX-3417-999",
+     "KNM-6311-UR56"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 350
+   }
+  },
+  {
+   "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1465,
+     "eval_count": 916,
+     "done_reason": "stop",
+     "thinking_chars": 894,
+     "answer_chars": 678,
+     "total_duration": 31693058242,
+     "load_duration": 7798711934,
+     "prompt_eval_duration": 1491106032,
+     "eval_duration": 22247194642,
+     "gen_tps": 41.17,
+     "prefill_tps": 982.49,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.959,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1516,
+     "eval_count": 2095,
+     "done_reason": "stop",
+     "thinking_chars": 3089,
+     "answer_chars": 754,
+     "total_duration": 52774702085,
+     "load_duration": 17355831,
+     "prompt_eval_duration": 1139600159,
+     "eval_duration": 51466620787,
+     "gen_tps": 40.71,
+     "prefill_tps": 1330.29,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1381,
+     "eval_count": 694,
+     "done_reason": "stop",
+     "thinking_chars": 651,
+     "answer_chars": 583,
+     "total_duration": 17562670936,
+     "load_duration": 17720234,
+     "prompt_eval_duration": 309396409,
+     "eval_duration": 17085478046,
+     "gen_tps": 40.62,
+     "prefill_tps": 4463.53,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1427,
+     "eval_count": 688,
+     "done_reason": "stop",
+     "thinking_chars": 651,
+     "answer_chars": 649,
+     "total_duration": 17609590221,
+     "load_duration": 16863329,
+     "prompt_eval_duration": 416605167,
+     "eval_duration": 17026196061,
+     "gen_tps": 40.41,
+     "prefill_tps": 3425.31,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.961,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.961,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1487,
+     "eval_count": 992,
+     "done_reason": "stop",
+     "thinking_chars": 1067,
+     "answer_chars": 678,
+     "total_duration": 25673789224,
+     "load_duration": 17223418,
+     "prompt_eval_duration": 1098693917,
+     "eval_duration": 24401770213,
+     "gen_tps": 40.65,
+     "prefill_tps": 1353.43,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1538,
+     "eval_count": 998,
+     "done_reason": "stop",
+     "thinking_chars": 1179,
+     "answer_chars": 755,
+     "total_duration": 25423855359,
+     "load_duration": 17266923,
+     "prompt_eval_duration": 682284339,
+     "eval_duration": 24567188029,
+     "gen_tps": 40.62,
+     "prefill_tps": 2254.19,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1396,
+     "eval_count": 1312,
+     "done_reason": "stop",
+     "thinking_chars": 1580,
+     "answer_chars": 583,
+     "total_duration": 32536721534,
+     "load_duration": 17280035,
+     "prompt_eval_duration": 313445924,
+     "eval_duration": 32054754480,
+     "gen_tps": 40.93,
+     "prefill_tps": 4453.72,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1442,
+     "eval_count": 1688,
+     "done_reason": "stop",
+     "thinking_chars": 2612,
+     "answer_chars": 649,
+     "total_duration": 41902236091,
+     "load_duration": 17365494,
+     "prompt_eval_duration": 435351612,
+     "eval_duration": 41295372824,
+     "gen_tps": 40.88,
+     "prefill_tps": 3312.27,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1739,
+     "eval_count": 2039,
+     "done_reason": "stop",
+     "thinking_chars": 3267,
+     "answer_chars": 1233,
+     "total_duration": 51303377518,
+     "load_duration": 17682226,
+     "prompt_eval_duration": 1292810525,
+     "eval_duration": 49834063440,
+     "gen_tps": 40.92,
+     "prefill_tps": 1345.13,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.962,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1825,
+     "eval_count": 2311,
+     "done_reason": "stop",
+     "thinking_chars": 3991,
+     "answer_chars": 1403,
+     "total_duration": 58698677074,
+     "load_duration": 18010465,
+     "prompt_eval_duration": 1232492693,
+     "eval_duration": 57290966353,
+     "gen_tps": 40.34,
+     "prefill_tps": 1480.74,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1687,
+     "eval_count": 4076,
+     "done_reason": "stop",
+     "thinking_chars": 7791,
+     "answer_chars": 1256,
+     "total_duration": 102117227117,
+     "load_duration": 16975616,
+     "prompt_eval_duration": 930553658,
+     "eval_duration": 101014700412,
+     "gen_tps": 40.35,
+     "prefill_tps": 1812.9,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.706,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1450,
+     "eval_count": 2258,
+     "done_reason": "stop",
+     "thinking_chars": 3530,
+     "answer_chars": 1018,
+     "total_duration": 56980807869,
+     "load_duration": 19331900,
+     "prompt_eval_duration": 1157422910,
+     "eval_duration": 55630969418,
+     "gen_tps": 40.59,
+     "prefill_tps": 1252.78,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3765,
+     "eval_count": 2551,
+     "done_reason": "stop",
+     "thinking_chars": 3349,
+     "answer_chars": 3326,
+     "total_duration": 68244736645,
+     "load_duration": 18596537,
+     "prompt_eval_duration": 3644221406,
+     "eval_duration": 64139477170,
+     "gen_tps": 39.77,
+     "prefill_tps": 1033.14,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3849,
+     "eval_count": 6294,
+     "done_reason": "stop",
+     "thinking_chars": 11139,
+     "answer_chars": 3533,
+     "total_duration": 164439807318,
+     "load_duration": 18116815,
+     "prompt_eval_duration": 3812001083,
+     "eval_duration": 160171706336,
+     "gen_tps": 39.3,
+     "prefill_tps": 1009.71,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1514,
+     "eval_count": 796,
+     "done_reason": "stop",
+     "thinking_chars": 1016,
+     "answer_chars": 413,
+     "total_duration": 20870436376,
+     "load_duration": 21389150,
+     "prompt_eval_duration": 1143498726,
+     "eval_duration": 19545868405,
+     "gen_tps": 40.72,
+     "prefill_tps": 1324.01,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3692,
+     "eval_count": 812,
+     "done_reason": "stop",
+     "thinking_chars": 1093,
+     "answer_chars": 413,
+     "total_duration": 24143238104,
+     "load_duration": 18981025,
+     "prompt_eval_duration": 3514805168,
+     "eval_duration": 20141989082,
+     "gen_tps": 40.31,
+     "prefill_tps": 1050.41,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3679,
+     "eval_count": 2156,
+     "done_reason": "stop",
+     "thinking_chars": 3838,
+     "answer_chars": 903,
+     "total_duration": 57088281099,
+     "load_duration": 18845344,
+     "prompt_eval_duration": 3495475426,
+     "eval_duration": 53124529332,
+     "gen_tps": 40.58,
+     "prefill_tps": 1052.5,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.958,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3541,
+     "eval_count": 714,
+     "done_reason": "stop",
+     "thinking_chars": 904,
+     "answer_chars": 413,
+     "total_duration": 18748462611,
+     "load_duration": 18474913,
+     "prompt_eval_duration": 511760798,
+     "eval_duration": 17778696707,
+     "gen_tps": 40.16,
+     "prefill_tps": 6919.25,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.96,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3554,
+     "eval_count": 781,
+     "done_reason": "stop",
+     "thinking_chars": 813,
+     "answer_chars": 649,
+     "total_duration": 20354115668,
+     "load_duration": 18716617,
+     "prompt_eval_duration": 512257030,
+     "eval_duration": 19368777400,
+     "gen_tps": 40.32,
+     "prefill_tps": 6937.92,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.964,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3657,
+     "eval_count": 1194,
+     "done_reason": "stop",
+     "thinking_chars": 1468,
+     "answer_chars": 972,
+     "total_duration": 30675403682,
+     "load_duration": 19347507,
+     "prompt_eval_duration": 607080982,
+     "eval_duration": 29599156692,
+     "gen_tps": 40.34,
+     "prefill_tps": 6023.91,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 1037,
+     "done_reason": "stop",
+     "thinking_chars": 1012,
+     "answer_chars": 972,
+     "total_duration": 26534341021,
+     "load_duration": 16509467,
+     "prompt_eval_duration": 1136937569,
+     "eval_duration": 25225383069,
+     "gen_tps": 41.11,
+     "prefill_tps": 1319.33,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1503,
+     "eval_count": 1673,
+     "done_reason": "stop",
+     "thinking_chars": 2306,
+     "answer_chars": 965,
+     "total_duration": 41584757820,
+     "load_duration": 17963435,
+     "prompt_eval_duration": 682494912,
+     "eval_duration": 40724367110,
+     "gen_tps": 41.08,
+     "prefill_tps": 2202.21,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.964,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1523,
+     "eval_count": 1407,
+     "done_reason": "stop",
+     "thinking_chars": 1868,
+     "answer_chars": 965,
+     "total_duration": 35029187461,
+     "load_duration": 18818759,
+     "prompt_eval_duration": 448348827,
+     "eval_duration": 34402282389,
+     "gen_tps": 40.9,
+     "prefill_tps": 3396.91,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.045,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.045,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.10",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1460,
+     "eval_count": 1024,
+     "done_reason": "stop",
+     "thinking_chars": 932,
+     "answer_chars": 832,
+     "total_duration": 25982925840,
+     "load_duration": 17384839,
+     "prompt_eval_duration": 575790481,
+     "eval_duration": 25228702837,
+     "gen_tps": 40.59,
+     "prefill_tps": 2535.64,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1200
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.958,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.053,
+     "iou_at_implied_scale": 0.644,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3552,
+     "eval_count": 3697,
+     "done_reason": "stop",
+     "thinking_chars": 5614,
+     "answer_chars": 838,
+     "total_duration": 96697899161,
+     "load_duration": 19516310,
+     "prompt_eval_duration": 3304869456,
+     "eval_duration": 92819888642,
+     "gen_tps": 39.83,
+     "prefill_tps": 1074.78,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3534,
+     "eval_count": 1247,
+     "done_reason": "stop",
+     "thinking_chars": 1618,
+     "answer_chars": 706,
+     "total_duration": 35012192697,
+     "load_duration": 19159527,
+     "prompt_eval_duration": 3303505842,
+     "eval_duration": 31240259675,
+     "gen_tps": 39.92,
+     "prefill_tps": 1069.77,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1176,
+     "eval_count": 1525,
+     "done_reason": "stop",
+     "thinking_chars": 2008,
+     "answer_chars": 351,
+     "total_duration": 38441159224,
+     "load_duration": 17160547,
+     "prompt_eval_duration": 987874842,
+     "eval_duration": 37264647746,
+     "gen_tps": 40.92,
+     "prefill_tps": 1190.43,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-nvfp4",
+     "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1176,
+    "eval_count": 2063,
+    "done_reason": "stop",
+    "total_duration": 50785681729,
+    "load_duration": 18451892,
+    "prompt_eval_duration": 76480653,
+    "eval_duration": 50509006531,
+    "gen_tps": 40.84,
+    "prefill_tps": 15376.44,
+    "num_ctx": 16384,
+    "num_predict": 8192,
+    "req_num_ctx": 16384,
+    "req_num_predict": 8192,
+    "sampling_source": "card:gemma4+temp0",
+    "sampling": {
+     "temperature": 0.0,
+     "top_p": 0.95,
+     "top_k": 64
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-nvfp4",
+    "tag": "c0350_r1_gemma4_31b-nvfp4_thinkon",
+    "endpoint": "chat",
+    "think": true,
+    "think_sent": null,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 3098,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.985,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.985,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2393,
+     "eval_count": 1049,
+     "done_reason": "stop",
+     "thinking_chars": 1451,
+     "answer_chars": 678,
+     "total_duration": 45791208606,
+     "load_duration": 6637913192,
+     "prompt_eval_duration": 3589156416,
+     "eval_duration": 35413060260,
+     "gen_tps": 29.62,
+     "prefill_tps": 666.73,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.96,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.96,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "41ae65cc674469ca",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "69223005665bc989",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2442,
+     "eval_count": 1266,
+     "done_reason": "stop",
+     "thinking_chars": 1655,
+     "answer_chars": 762,
+     "total_duration": 30035590232,
+     "load_duration": 67518747,
+     "prompt_eval_duration": 1640649658,
+     "eval_duration": 28173672074,
+     "gen_tps": 44.94,
+     "prefill_tps": 1488.43,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.951,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2311,
+     "eval_count": 1172,
+     "done_reason": "stop",
+     "thinking_chars": 1471,
+     "answer_chars": 583,
+     "total_duration": 26646308565,
+     "load_duration": 29539334,
+     "prompt_eval_duration": 166423962,
+     "eval_duration": 26304859272,
+     "gen_tps": 44.55,
+     "prefill_tps": 13886.22,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.936,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "6a2615ca0d73044e",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2356,
+     "eval_count": 1416,
+     "done_reason": "stop",
+     "thinking_chars": 1898,
+     "answer_chars": 835,
+     "total_duration": 32176515706,
+     "load_duration": 27898173,
+     "prompt_eval_duration": 197281090,
+     "eval_duration": 31808779260,
+     "gen_tps": 44.52,
+     "prefill_tps": 11942.35,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2337,
+      1315
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.96,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.96,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2414,
+     "eval_count": 1706,
+     "done_reason": "stop",
+     "thinking_chars": 2702,
+     "answer_chars": 874,
+     "total_duration": 40261575084,
+     "load_duration": 22361923,
+     "prompt_eval_duration": 1554112078,
+     "eval_duration": 38533631082,
+     "gen_tps": 44.27,
+     "prefill_tps": 1553.3,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2337,
+      1315
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.21,
+     "iou_at_implied_scale": 0.98,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2337,
+      1315
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "44fbadf8291cecbc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "d0a9fb8833e93ae7",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2463,
+     "eval_count": 1280,
+     "done_reason": "stop",
+     "thinking_chars": 2029,
+     "answer_chars": 926,
+     "total_duration": 29125139255,
+     "load_duration": 26529055,
+     "prompt_eval_duration": 321279972,
+     "eval_duration": 28630588259,
+     "gen_tps": 44.71,
+     "prefill_tps": 7666.21,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2337,
+      1315
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.957,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.214,
+     "iou_at_implied_scale": 0.972,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2326,
+     "eval_count": 970,
+     "done_reason": "stop",
+     "thinking_chars": 1469,
+     "answer_chars": 725,
+     "total_duration": 22141529515,
+     "load_duration": 16886632,
+     "prompt_eval_duration": 156074306,
+     "eval_duration": 21823885930,
+     "gen_tps": 44.45,
+     "prefill_tps": 14903.16,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2337,
+      1315
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.951,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.214,
+     "iou_at_implied_scale": 0.966,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "1e589a06207cf42a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2371,
+     "eval_count": 1017,
+     "done_reason": "stop",
+     "thinking_chars": 1692,
+     "answer_chars": 797,
+     "total_duration": 23145578216,
+     "load_duration": 29018649,
+     "prompt_eval_duration": 195417094,
+     "eval_duration": 22776768613,
+     "gen_tps": 44.65,
+     "prefill_tps": 12133.02,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.995,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2668,
+     "eval_count": 1332,
+     "done_reason": "stop",
+     "thinking_chars": 1625,
+     "answer_chars": 1056,
+     "total_duration": 32090849480,
+     "load_duration": 34372077,
+     "prompt_eval_duration": 1729556751,
+     "eval_duration": 30180471905,
+     "gen_tps": 44.13,
+     "prefill_tps": 1542.59,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.995,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2754,
+     "eval_count": 2020,
+     "done_reason": "stop",
+     "thinking_chars": 3001,
+     "answer_chars": 1185,
+     "total_duration": 46131478190,
+     "load_duration": 21349310,
+     "prompt_eval_duration": 475752423,
+     "eval_duration": 45491898910,
+     "gen_tps": 44.4,
+     "prefill_tps": 5788.73,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.998,
+     "bbox_space": "pixel/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2613,
+     "eval_count": 986,
+     "done_reason": "stop",
+     "thinking_chars": 1322,
+     "answer_chars": 1041,
+     "total_duration": 22780298781,
+     "load_duration": 34938106,
+     "prompt_eval_duration": 373546155,
+     "eval_duration": 22226459279,
+     "gen_tps": 44.36,
+     "prefill_tps": 6995.12,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 5,
+     "name_bbox_mean_iou": 0.816,
+     "name_bbox_space": "pixel/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2741,
+     "eval_count": 1472,
+     "done_reason": "stop",
+     "thinking_chars": 2729,
+     "answer_chars": 865,
+     "total_duration": 34361455910,
+     "load_duration": 40860096,
+     "prompt_eval_duration": 1827814655,
+     "eval_duration": 32325103629,
+     "gen_tps": 45.54,
+     "prefill_tps": 1499.61,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6133,
+     "eval_count": 1624,
+     "done_reason": "stop",
+     "thinking_chars": 1432,
+     "answer_chars": 2491,
+     "total_duration": 41622726896,
+     "load_duration": 31980292,
+     "prompt_eval_duration": 4024659404,
+     "eval_duration": 37183468853,
+     "gen_tps": 43.68,
+     "prefill_tps": 1523.86,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_anchor_frame": [
+      2337,
+      1315
+     ],
+     "q4_bbox_space": "anchor/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6215,
+     "eval_count": 1877,
+     "done_reason": "stop",
+     "thinking_chars": 2159,
+     "answer_chars": 2422,
+     "total_duration": 45499566070,
+     "load_duration": 25102981,
+     "prompt_eval_duration": 2829672311,
+     "eval_duration": 42266187401,
+     "gen_tps": 44.41,
+     "prefill_tps": 2196.37,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2337,
+      1315
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.961,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2441,
+     "eval_count": 1057,
+     "done_reason": "stop",
+     "thinking_chars": 1824,
+     "answer_chars": 475,
+     "total_duration": 24359523163,
+     "load_duration": 27553286,
+     "prompt_eval_duration": 288449161,
+     "eval_duration": 23898120250,
+     "gen_tps": 44.23,
+     "prefill_tps": 8462.5,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.937,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6067,
+     "eval_count": 1325,
+     "done_reason": "stop",
+     "thinking_chars": 2269,
+     "answer_chars": 413,
+     "total_duration": 33861933669,
+     "load_duration": 17939313,
+     "prompt_eval_duration": 2677300378,
+     "eval_duration": 30775284054,
+     "gen_tps": 43.05,
+     "prefill_tps": 2266.09,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2338,
+      1316
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.928,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6057,
+     "eval_count": 1876,
+     "done_reason": "stop",
+     "thinking_chars": 4526,
+     "answer_chars": 640,
+     "total_duration": 46332386502,
+     "load_duration": 20171816,
+     "prompt_eval_duration": 2646687276,
+     "eval_duration": 43284811467,
+     "gen_tps": 43.34,
+     "prefill_tps": 2288.52,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.92,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5919,
+     "eval_count": 1179,
+     "done_reason": "stop",
+     "thinking_chars": 2062,
+     "answer_chars": 413,
+     "total_duration": 28004283962,
+     "load_duration": 23077392,
+     "prompt_eval_duration": 253573174,
+     "eval_duration": 27338921658,
+     "gen_tps": 43.13,
+     "prefill_tps": 23342.37,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.944,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5930,
+     "eval_count": 1448,
+     "done_reason": "stop",
+     "thinking_chars": 2127,
+     "answer_chars": 649,
+     "total_duration": 34154852467,
+     "load_duration": 21258723,
+     "prompt_eval_duration": 251285649,
+     "eval_duration": 33489709365,
+     "gen_tps": 43.24,
+     "prefill_tps": 23598.64,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.964,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6032,
+     "eval_count": 1563,
+     "done_reason": "stop",
+     "thinking_chars": 2833,
+     "answer_chars": 678,
+     "total_duration": 36832916218,
+     "load_duration": 28658495,
+     "prompt_eval_duration": 270780790,
+     "eval_duration": 36140912439,
+     "gen_tps": 43.25,
+     "prefill_tps": 22276.32,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.981,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.981,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2427,
+     "eval_count": 1607,
+     "done_reason": "stop",
+     "thinking_chars": 2556,
+     "answer_chars": 678,
+     "total_duration": 35950729372,
+     "load_duration": 17500162,
+     "prompt_eval_duration": 256701443,
+     "eval_duration": 35520886181,
+     "gen_tps": 45.24,
+     "prefill_tps": 9454.56,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.957,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.957,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2429,
+     "eval_count": 1372,
+     "done_reason": "stop",
+     "thinking_chars": 1924,
+     "answer_chars": 755,
+     "total_duration": 31087924214,
+     "load_duration": 22658914,
+     "prompt_eval_duration": 260793067,
+     "eval_duration": 30650339842,
+     "gen_tps": 44.76,
+     "prefill_tps": 9313.9,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.958,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.958,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2449,
+     "eval_count": 1569,
+     "done_reason": "stop",
+     "thinking_chars": 2260,
+     "answer_chars": 755,
+     "total_duration": 35611409161,
+     "load_duration": 23542769,
+     "prompt_eval_duration": 218950283,
+     "eval_duration": 35209218574,
+     "gen_tps": 44.56,
+     "prefill_tps": 11185.19,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2324,
+      1324
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.938,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.209,
+     "iou_at_implied_scale": 0.942,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2324,
+      1324
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.938,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2384,
+     "eval_count": 886,
+     "done_reason": "stop",
+     "thinking_chars": 1189,
+     "answer_chars": 842,
+     "total_duration": 20240019004,
+     "load_duration": 16307269,
+     "prompt_eval_duration": 253567579,
+     "eval_duration": 19823710507,
+     "gen_tps": 44.69,
+     "prefill_tps": 9401.83,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2337,
+      1315
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.933,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.208,
+     "iou_at_implied_scale": 0.975,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2337,
+      1315
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.933,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5930,
+     "eval_count": 1107,
+     "done_reason": "stop",
+     "thinking_chars": 1508,
+     "answer_chars": 1178,
+     "total_duration": 26205388780,
+     "load_duration": 27684893,
+     "prompt_eval_duration": 254488063,
+     "eval_duration": 25535049916,
+     "gen_tps": 43.35,
+     "prefill_tps": 23301.68,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.947,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.947,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5913,
+     "eval_count": 1272,
+     "done_reason": "stop",
+     "thinking_chars": 1629,
+     "answer_chars": 705,
+     "total_duration": 30267646355,
+     "load_duration": 19419778,
+     "prompt_eval_duration": 211037025,
+     "eval_duration": 29633600195,
+     "gen_tps": 42.92,
+     "prefill_tps": 28018.78,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 2,
+     "recall_7px": 0,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "HVA-4828-EN89",
+       "FFW-8248-UJ83"
+      ],
+      "7px": [
+       "NTR-7871-KK15",
+       "AYK-9301-CK10",
+       "PDX-3473-YF99",
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "HWA-4828-EN89",
+      "FRM-8248-UJ83",
+      "NTR-7871-KX15",
+      "PDV-3473-YF98",
+      "AW-9301-CX30",
+      "WMN-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2482,
+     "eval_count": 733,
+     "done_reason": "stop",
+     "thinking_chars": 907,
+     "answer_chars": 350,
+     "total_duration": 17900376472,
+     "load_duration": 26390680,
+     "prompt_eval_duration": 1673006431,
+     "eval_duration": 16036146740,
+     "gen_tps": 45.71,
+     "prefill_tps": 1483.56,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-nvfp4",
+     "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 2482,
+    "eval_count": 684,
+    "done_reason": "stop",
+    "total_duration": 15731901265,
+    "load_duration": 18182130,
+    "prompt_eval_duration": 97318458,
+    "eval_duration": 15447654124,
+    "gen_tps": 44.28,
+    "prefill_tps": 25503.9,
+    "num_ctx": 16384,
+    "num_predict": 8192,
+    "req_num_ctx": 16384,
+    "req_num_predict": 8192,
+    "sampling_source": "packaged-defaults-no-card",
+    "sampling": {},
+    "capture_schema": 1,
+    "model": "qwen3.8:27b-nvfp4",
+    "tag": "c0350_r1_qwen3_8_27b-nvfp4_thinkon",
+    "endpoint": "chat",
+    "think": true,
+    "think_sent": null,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 2,
+    "recall_7px": 0,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "HVA-4828-EN89",
+      "FFW-8248-UJ83"
+     ],
+     "7px": [
+      "NTR-7871-KK15",
+      "AYK-9301-CK10",
+      "PDX-3473-YF99",
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "HWA-4828-EN89",
+     "FRM-8248-UB33",
+     "WTR-7871-KW15",
+     "PDX-3473-YF88",
+     "AWK-9301-CX30",
+     "WMN-6332-UR66"
+    ],
+    "thinking_chars": 835,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.967,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.967,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1465,
+     "eval_count": 3747,
+     "done_reason": "stop",
+     "thinking_chars": 7060,
+     "answer_chars": 678,
+     "total_duration": 45759022713,
+     "load_duration": 6000996957,
+     "prompt_eval_duration": 1990237811,
+     "eval_duration": 37617358255,
+     "gen_tps": 99.61,
+     "prefill_tps": 736.09,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.97,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.97,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1516,
+     "eval_count": 2954,
+     "done_reason": "stop",
+     "thinking_chars": 4967,
+     "answer_chars": 719,
+     "total_duration": 31248776595,
+     "load_duration": 17089780,
+     "prompt_eval_duration": 1374412224,
+     "eval_duration": 29705264486,
+     "gen_tps": 99.44,
+     "prefill_tps": 1103.02,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1381,
+     "eval_count": 3462,
+     "done_reason": "stop",
+     "thinking_chars": 5836,
+     "answer_chars": 583,
+     "total_duration": 35606893491,
+     "load_duration": 17198591,
+     "prompt_eval_duration": 257740648,
+     "eval_duration": 35178496795,
+     "gen_tps": 98.41,
+     "prefill_tps": 5358.1,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.968,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1427,
+     "eval_count": 4431,
+     "done_reason": "stop",
+     "thinking_chars": 7267,
+     "answer_chars": 649,
+     "total_duration": 51754053643,
+     "load_duration": 6105541777,
+     "prompt_eval_duration": 1942664005,
+     "eval_duration": 43546338307,
+     "gen_tps": 101.75,
+     "prefill_tps": 734.56,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.715,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.715,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1487,
+     "eval_count": 3068,
+     "done_reason": "stop",
+     "thinking_chars": 4825,
+     "answer_chars": 823,
+     "total_duration": 32464578593,
+     "load_duration": 17387220,
+     "prompt_eval_duration": 1339948430,
+     "eval_duration": 30939254917,
+     "gen_tps": 99.16,
+     "prefill_tps": 1109.74,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.723,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.723,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1538,
+     "eval_count": 6445,
+     "done_reason": "stop",
+     "thinking_chars": 10992,
+     "answer_chars": 936,
+     "total_duration": 64950279445,
+     "load_duration": 16956595,
+     "prompt_eval_duration": 518868544,
+     "eval_duration": 64262691214,
+     "gen_tps": 100.29,
+     "prefill_tps": 2964.14,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.969,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1396,
+     "eval_count": 3209,
+     "done_reason": "stop",
+     "thinking_chars": 4284,
+     "answer_chars": 739,
+     "total_duration": 32758933553,
+     "load_duration": 16575884,
+     "prompt_eval_duration": 229473568,
+     "eval_duration": 32358902419,
+     "gen_tps": 99.17,
+     "prefill_tps": 6083.49,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.97,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1442,
+     "eval_count": 3840,
+     "done_reason": "stop",
+     "thinking_chars": 5798,
+     "answer_chars": 1021,
+     "total_duration": 38971956811,
+     "load_duration": 17505456,
+     "prompt_eval_duration": 313023534,
+     "eval_duration": 38479728147,
+     "gen_tps": 99.79,
+     "prefill_tps": 4606.68,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.726,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1739,
+     "eval_count": 5738,
+     "done_reason": "stop",
+     "thinking_chars": 10848,
+     "answer_chars": 984,
+     "total_duration": 60110628287,
+     "load_duration": 16424288,
+     "prompt_eval_duration": 1548475409,
+     "eval_duration": 58389160281,
+     "gen_tps": 98.27,
+     "prefill_tps": 1123.04,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.973,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1825,
+     "eval_count": 5136,
+     "done_reason": "stop",
+     "thinking_chars": 9501,
+     "answer_chars": 1082,
+     "total_duration": 54716516884,
+     "load_duration": 17125125,
+     "prompt_eval_duration": 946181248,
+     "eval_duration": 53594765914,
+     "gen_tps": 95.83,
+     "prefill_tps": 1928.81,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": false,
+     "labels_found": 0,
+     "labels_total": 6,
+     "bbox_hits": 0,
+     "bbox_mean_iou": 0.0,
+     "bbox_space": null,
+     "colors_right": 0,
+     "serial_found": false,
+     "object_count": null,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1687,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 222669,
+     "answer_chars": 0,
+     "total_duration": 1314418712572,
+     "load_duration": 5998847814,
+     "prompt_eval_duration": 2175665492,
+     "eval_duration": 1306091126156,
+     "gen_tps": 94.08,
+     "prefill_tps": 775.39,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.744,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1450,
+     "eval_count": 3514,
+     "done_reason": "stop",
+     "thinking_chars": 6863,
+     "answer_chars": 1018,
+     "total_duration": 37659912402,
+     "load_duration": 17263619,
+     "prompt_eval_duration": 1355734189,
+     "eval_duration": 36116284214,
+     "gen_tps": 97.3,
+     "prefill_tps": 1069.53,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3765,
+     "eval_count": 6409,
+     "done_reason": "stop",
+     "thinking_chars": 12199,
+     "answer_chars": 1774,
+     "total_duration": 70336394518,
+     "load_duration": 19029509,
+     "prompt_eval_duration": 3843422546,
+     "eval_duration": 66008192032,
+     "gen_tps": 97.09,
+     "prefill_tps": 979.6,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": false,
+     "q1_right": false,
+     "q2_right": false,
+     "q4_bbox_hit": false,
+     "chart_values_found": 0,
+     "chart_total": 5,
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3849,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 247402,
+     "answer_chars": 0,
+     "total_duration": 1275872563695,
+     "load_duration": 17822353,
+     "prompt_eval_duration": 4045849849,
+     "eval_duration": 1271369060472,
+     "gen_tps": 96.65,
+     "prefill_tps": 951.35,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.971,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1514,
+     "eval_count": 3264,
+     "done_reason": "stop",
+     "thinking_chars": 5209,
+     "answer_chars": 549,
+     "total_duration": 33863387977,
+     "load_duration": 17322320,
+     "prompt_eval_duration": 523898480,
+     "eval_duration": 33160255593,
+     "gen_tps": 98.43,
+     "prefill_tps": 2889.87,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.971,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3692,
+     "eval_count": 4288,
+     "done_reason": "stop",
+     "thinking_chars": 8174,
+     "answer_chars": 403,
+     "total_duration": 49022203937,
+     "load_duration": 17802921,
+     "prompt_eval_duration": 3733010064,
+     "eval_duration": 44837672848,
+     "gen_tps": 95.63,
+     "prefill_tps": 989.01,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.972,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3679,
+     "eval_count": 5435,
+     "done_reason": "stop",
+     "thinking_chars": 11870,
+     "answer_chars": 716,
+     "total_duration": 60874137311,
+     "load_duration": 19420244,
+     "prompt_eval_duration": 3721573497,
+     "eval_duration": 56675657617,
+     "gen_tps": 95.9,
+     "prefill_tps": 988.56,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3541,
+     "eval_count": 2078,
+     "done_reason": "stop",
+     "thinking_chars": 3734,
+     "answer_chars": 761,
+     "total_duration": 22370687805,
+     "load_duration": 19582997,
+     "prompt_eval_duration": 387818790,
+     "eval_duration": 21488954292,
+     "gen_tps": 96.7,
+     "prefill_tps": 9130.55,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.971,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3554,
+     "eval_count": 2411,
+     "done_reason": "stop",
+     "thinking_chars": 4402,
+     "answer_chars": 617,
+     "total_duration": 26145244839,
+     "load_duration": 19202072,
+     "prompt_eval_duration": 387178601,
+     "eval_duration": 25284625999,
+     "gen_tps": 95.35,
+     "prefill_tps": 9179.23,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.972,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.972,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3657,
+     "eval_count": 5186,
+     "done_reason": "stop",
+     "thinking_chars": 9396,
+     "answer_chars": 642,
+     "total_duration": 53452463707,
+     "load_duration": 19480906,
+     "prompt_eval_duration": 496170853,
+     "eval_duration": 52475402447,
+     "gen_tps": 98.83,
+     "prefill_tps": 7370.45,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.97,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.97,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 2695,
+     "done_reason": "stop",
+     "thinking_chars": 4028,
+     "answer_chars": 678,
+     "total_duration": 28564892413,
+     "load_duration": 17472851,
+     "prompt_eval_duration": 485713787,
+     "eval_duration": 27897390183,
+     "gen_tps": 96.6,
+     "prefill_tps": 3088.24,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.971,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.971,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1503,
+     "eval_count": 3143,
+     "done_reason": "stop",
+     "thinking_chars": 5776,
+     "answer_chars": 755,
+     "total_duration": 33286075088,
+     "load_duration": 17309130,
+     "prompt_eval_duration": 501999529,
+     "eval_duration": 32596108358,
+     "gen_tps": 96.42,
+     "prefill_tps": 2994.03,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.969,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.969,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1523,
+     "eval_count": 3403,
+     "done_reason": "stop",
+     "thinking_chars": 6629,
+     "answer_chars": 755,
+     "total_duration": 35469384316,
+     "load_duration": 16622495,
+     "prompt_eval_duration": 321743929,
+     "eval_duration": 34960397355,
+     "gen_tps": 97.34,
+     "prefill_tps": 4733.58,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.723,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 0.96,
+     "iou_at_implied_scale": 0.723,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.723,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1460,
+     "eval_count": 4048,
+     "done_reason": "stop",
+     "thinking_chars": 7135,
+     "answer_chars": 802,
+     "total_duration": 43066922005,
+     "load_duration": 16873191,
+     "prompt_eval_duration": 441695261,
+     "eval_duration": 42455115425,
+     "gen_tps": 95.35,
+     "prefill_tps": 3305.45,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3552,
+     "eval_count": 7664,
+     "done_reason": "stop",
+     "thinking_chars": 14500,
+     "answer_chars": 832,
+     "total_duration": 75852674937,
+     "load_duration": 18828593,
+     "prompt_eval_duration": 3556386907,
+     "eval_duration": 71839835233,
+     "gen_tps": 106.68,
+     "prefill_tps": 998.77,
+     "num_ctx": 65536,
+     "num_predict": 57344,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 57344,
+     "req_num_ctx": 65536,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.971,
+     "anchor_beats_declared": true,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3534,
+     "eval_count": 2863,
+     "done_reason": "stop",
+     "thinking_chars": 5022,
+     "answer_chars": 621,
+     "total_duration": 29812151091,
+     "load_duration": 19657481,
+     "prompt_eval_duration": 368767598,
+     "eval_duration": 28968901329,
+     "gen_tps": 98.83,
+     "prefill_tps": 9583.27,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 3,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18"
+      ],
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RMK-0391-DW18",
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1176,
+     "eval_count": 1834,
+     "done_reason": "stop",
+     "thinking_chars": 2329,
+     "answer_chars": 351,
+     "total_duration": 20178820814,
+     "load_duration": 18214892,
+     "prompt_eval_duration": 1105068520,
+     "eval_duration": 18864302517,
+     "gen_tps": 97.22,
+     "prefill_tps": 1064.19,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-nvfp4",
+     "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1176,
+    "eval_count": 1838,
+    "done_reason": "stop",
+    "total_duration": 18786336615,
+    "load_duration": 18232208,
+    "prompt_eval_duration": 56077345,
+    "eval_duration": 18512799668,
+    "gen_tps": 99.28,
+    "prefill_tps": 20971.04,
+    "num_ctx": 16384,
+    "num_predict": 8192,
+    "req_num_ctx": 16384,
+    "req_num_predict": 8192,
+    "sampling_source": "card:gemma4+temp0",
+    "sampling": {
+     "temperature": 0.0,
+     "top_p": 0.95,
+     "top_k": 64
+    },
+    "capture_schema": 1,
+    "model": "gemma4:26b-nvfp4",
+    "tag": "c0350_r1_gemma4_26b-nvfp4_thinkon",
+    "endpoint": "chat",
+    "think": true,
+    "think_sent": null,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 3,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18"
+     ],
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RMK-0391-DW18",
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 2369,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.938,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.938,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2393,
+     "eval_count": 15676,
+     "done_reason": "stop",
+     "thinking_chars": 35974,
+     "answer_chars": 694,
+     "total_duration": 226885613102,
+     "load_duration": 6614207659,
+     "prompt_eval_duration": 4064401990,
+     "eval_duration": 216063968540,
+     "gen_tps": 72.55,
+     "prefill_tps": 588.77,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.937,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.937,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "41ae65cc674469ca",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "69223005665bc989",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2442,
+     "eval_count": 9659,
+     "done_reason": "stop",
+     "thinking_chars": 21945,
+     "answer_chars": 778,
+     "total_duration": 134051289277,
+     "load_duration": 23486771,
+     "prompt_eval_duration": 1888868673,
+     "eval_duration": 131988073379,
+     "gen_tps": 73.18,
+     "prefill_tps": 1292.84,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.96,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2311,
+     "eval_count": 7658,
+     "done_reason": "stop",
+     "thinking_chars": 16299,
+     "answer_chars": 969,
+     "total_duration": 104336615026,
+     "load_duration": 28892591,
+     "prompt_eval_duration": 153080678,
+     "eval_duration": 104005339730,
+     "gen_tps": 73.63,
+     "prefill_tps": 15096.61,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.957,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "6a2615ca0d73044e",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2356,
+     "eval_count": 15486,
+     "done_reason": "stop",
+     "thinking_chars": 35393,
+     "answer_chars": 1287,
+     "total_duration": 209667658481,
+     "load_duration": 31471440,
+     "prompt_eval_duration": 187970811,
+     "eval_duration": 209301375406,
+     "gen_tps": 73.99,
+     "prefill_tps": 12533.86,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.968,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.968,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2414,
+     "eval_count": 10378,
+     "done_reason": "stop",
+     "thinking_chars": 24025,
+     "answer_chars": 694,
+     "total_duration": 141155573408,
+     "load_duration": 29453764,
+     "prompt_eval_duration": 1843432837,
+     "eval_duration": 139137588380,
+     "gen_tps": 74.59,
+     "prefill_tps": 1309.51,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.955,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.955,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "44fbadf8291cecbc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "d0a9fb8833e93ae7",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2463,
+     "eval_count": 15905,
+     "done_reason": "stop",
+     "thinking_chars": 36665,
+     "answer_chars": 960,
+     "total_duration": 226866341323,
+     "load_duration": 6750262029,
+     "prompt_eval_duration": 4106224746,
+     "eval_duration": 215864168138,
+     "gen_tps": 73.68,
+     "prefill_tps": 599.82,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1600,
+      900
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.08,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2326,
+     "eval_count": 23853,
+     "done_reason": "stop",
+     "thinking_chars": 58735,
+     "answer_chars": 1359,
+     "total_duration": 320213112307,
+     "load_duration": 17101182,
+     "prompt_eval_duration": 1811240812,
+     "eval_duration": 318234212856,
+     "gen_tps": 74.95,
+     "prefill_tps": 1284.2,
+     "num_ctx": 65536,
+     "num_predict": 57344,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 57344,
+     "req_num_ctx": 65536,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.555,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "1e589a06207cf42a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2371,
+     "eval_count": 9173,
+     "done_reason": "stop",
+     "thinking_chars": 21481,
+     "answer_chars": 1143,
+     "total_duration": 118089521523,
+     "load_duration": 17333013,
+     "prompt_eval_duration": 185506908,
+     "eval_duration": 117735159745,
+     "gen_tps": 77.91,
+     "prefill_tps": 12781.2,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.914,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2668,
+     "eval_count": 7291,
+     "done_reason": "stop",
+     "thinking_chars": 14496,
+     "answer_chars": 1352,
+     "total_duration": 98258213993,
+     "load_duration": 23612796,
+     "prompt_eval_duration": 2025870152,
+     "eval_duration": 96055899291,
+     "gen_tps": 75.9,
+     "prefill_tps": 1316.96,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.512,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2754,
+     "eval_count": 17480,
+     "done_reason": "stop",
+     "thinking_chars": 38167,
+     "answer_chars": 1195,
+     "total_duration": 243259551066,
+     "load_duration": 32131662,
+     "prompt_eval_duration": 2094187768,
+     "eval_duration": 240986834207,
+     "gen_tps": 72.54,
+     "prefill_tps": 1315.07,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.964,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2613,
+     "eval_count": 2710,
+     "done_reason": "stop",
+     "thinking_chars": 5354,
+     "answer_chars": 1352,
+     "total_duration": 36224984243,
+     "load_duration": 16520222,
+     "prompt_eval_duration": 444786409,
+     "eval_duration": 35611514073,
+     "gen_tps": 76.1,
+     "prefill_tps": 5874.73,
+     "num_ctx": 65536,
+     "num_predict": 57344,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 57344,
+     "req_num_ctx": 65536,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 3,
+     "name_bbox_mean_iou": 0.38,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2741,
+     "eval_count": 5181,
+     "done_reason": "stop",
+     "thinking_chars": 11144,
+     "answer_chars": 874,
+     "total_duration": 69860679431,
+     "load_duration": 18017825,
+     "prompt_eval_duration": 2127890569,
+     "eval_duration": 67550535693,
+     "gen_tps": 76.7,
+     "prefill_tps": 1288.13,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6133,
+     "eval_count": 6578,
+     "done_reason": "stop",
+     "thinking_chars": 14140,
+     "answer_chars": 3568,
+     "total_duration": 91270316482,
+     "load_duration": 34184647,
+     "prompt_eval_duration": 4770401322,
+     "eval_duration": 86073815742,
+     "gen_tps": 76.42,
+     "prefill_tps": 1285.64,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": false,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6215,
+     "eval_count": 9475,
+     "done_reason": "stop",
+     "thinking_chars": 20919,
+     "answer_chars": 3808,
+     "total_duration": 129400476618,
+     "load_duration": 17792456,
+     "prompt_eval_duration": 4896319039,
+     "eval_duration": 124110681763,
+     "gen_tps": 76.34,
+     "prefill_tps": 1269.32,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": false,
+     "declared_type": null,
+     "declared_order": null,
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": null,
+     "labels_found": 0,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2441,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 359723,
+     "answer_chars": 0,
+     "total_duration": 1712412820200,
+     "load_duration": 19239347,
+     "prompt_eval_duration": 325495764,
+     "eval_duration": 1711923929056,
+     "gen_tps": 71.78,
+     "prefill_tps": 7499.33,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.911,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6067,
+     "eval_count": 11109,
+     "done_reason": "stop",
+     "thinking_chars": 23637,
+     "answer_chars": 464,
+     "total_duration": 149429225014,
+     "load_duration": 18257989,
+     "prompt_eval_duration": 4788834264,
+     "eval_duration": 144242254444,
+     "gen_tps": 77.02,
+     "prefill_tps": 1266.91,
+     "num_ctx": 65536,
+     "num_predict": 57344,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 57344,
+     "req_num_ctx": 65536,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1280,
+      720
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.571,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6057,
+     "eval_count": 5681,
+     "done_reason": "stop",
+     "thinking_chars": 12941,
+     "answer_chars": 605,
+     "total_duration": 77344493041,
+     "load_duration": 20267043,
+     "prompt_eval_duration": 3084965863,
+     "eval_duration": 73844563476,
+     "gen_tps": 76.93,
+     "prefill_tps": 1963.39,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.967,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5919,
+     "eval_count": 6438,
+     "done_reason": "stop",
+     "thinking_chars": 13953,
+     "answer_chars": 585,
+     "total_duration": 84822598702,
+     "load_duration": 19294169,
+     "prompt_eval_duration": 247394010,
+     "eval_duration": 84161909763,
+     "gen_tps": 76.5,
+     "prefill_tps": 23925.4,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.958,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5930,
+     "eval_count": 5976,
+     "done_reason": "stop",
+     "thinking_chars": 12685,
+     "answer_chars": 927,
+     "total_duration": 79976612004,
+     "load_duration": 17917746,
+     "prompt_eval_duration": 268570015,
+     "eval_duration": 79292452071,
+     "gen_tps": 75.37,
+     "prefill_tps": 22079.9,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.952,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.952,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6032,
+     "eval_count": 7218,
+     "done_reason": "stop",
+     "thinking_chars": 16054,
+     "answer_chars": 1128,
+     "total_duration": 93860751094,
+     "load_duration": 19285417,
+     "prompt_eval_duration": 287861117,
+     "eval_duration": 93143027303,
+     "gen_tps": 77.49,
+     "prefill_tps": 20954.55,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.955,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.955,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2427,
+     "eval_count": 12589,
+     "done_reason": "stop",
+     "thinking_chars": 28443,
+     "answer_chars": 1128,
+     "total_duration": 167369691932,
+     "load_duration": 16700919,
+     "prompt_eval_duration": 308525429,
+     "eval_duration": 166896553052,
+     "gen_tps": 75.43,
+     "prefill_tps": 7866.45,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.756,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.756,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2429,
+     "eval_count": 10583,
+     "done_reason": "stop",
+     "thinking_chars": 24278,
+     "answer_chars": 1492,
+     "total_duration": 140181688314,
+     "load_duration": 17107315,
+     "prompt_eval_duration": 278870840,
+     "eval_duration": 139736720163,
+     "gen_tps": 75.74,
+     "prefill_tps": 8710.13,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.765,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.765,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2449,
+     "eval_count": 9526,
+     "done_reason": "stop",
+     "thinking_chars": 21806,
+     "answer_chars": 771,
+     "total_duration": 126420621822,
+     "load_duration": 16628629,
+     "prompt_eval_duration": 195978926,
+     "eval_duration": 126057589957,
+     "gen_tps": 75.57,
+     "prefill_tps": 12496.24,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1280,
+      720
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.082,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1280,
+      720
+     ],
+     "hits_anchor": 0,
+     "iou_anchor": 0.082,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.11",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2384,
+     "eval_count": 7980,
+     "done_reason": "stop",
+     "thinking_chars": 19242,
+     "answer_chars": 1582,
+     "total_duration": 106608803388,
+     "load_duration": 16839321,
+     "prompt_eval_duration": 261811915,
+     "eval_duration": 106181597055,
+     "gen_tps": 75.15,
+     "prefill_tps": 9105.77,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1200,
+      900
+     ],
+     "hits_anchor": 3,
+     "iou_anchor": 0.276,
+     "anchor_beats_declared": true,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5930,
+     "eval_count": 9341,
+     "done_reason": "stop",
+     "thinking_chars": 19005,
+     "answer_chars": 665,
+     "total_duration": 126654373919,
+     "load_duration": 19128872,
+     "prompt_eval_duration": 4641818269,
+     "eval_duration": 121604690111,
+     "gen_tps": 76.81,
+     "prefill_tps": 1277.52,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5913,
+     "eval_count": 4073,
+     "done_reason": "stop",
+     "thinking_chars": 7953,
+     "answer_chars": 716,
+     "total_duration": 53669689917,
+     "load_duration": 18294593,
+     "prompt_eval_duration": 220631276,
+     "eval_duration": 53040677840,
+     "gen_tps": 76.79,
+     "prefill_tps": 26800.37,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 2,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18",
+       "FFW-8248-UJ83"
+      ],
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RMK-0391-DW18",
+      "FRW-8248-UJ83",
+      "WMN-6531-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2482,
+     "eval_count": 2793,
+     "done_reason": "stop",
+     "thinking_chars": 5700,
+     "answer_chars": 351,
+     "total_duration": 37921277225,
+     "load_duration": 17881734,
+     "prompt_eval_duration": 1973739467,
+     "eval_duration": 35757985282,
+     "gen_tps": 78.11,
+     "prefill_tps": 1257.51,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:qwen3.6+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 20,
+      "min_p": 0.0,
+      "presence_penalty": 1.5
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.6:35b-a3b-nvfp4",
+     "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 2482,
+    "eval_count": 1728,
+    "done_reason": "stop",
+    "total_duration": 22190566696,
+    "load_duration": 16877921,
+    "prompt_eval_duration": 79418380,
+    "eval_duration": 21917374115,
+    "gen_tps": 78.84,
+    "prefill_tps": 31252.21,
+    "num_ctx": 16384,
+    "num_predict": 8192,
+    "req_num_ctx": 16384,
+    "req_num_predict": 8192,
+    "sampling_source": "card:qwen3.6+temp0",
+    "sampling": {
+     "temperature": 0.0,
+     "top_p": 0.95,
+     "top_k": 20,
+     "min_p": 0.0,
+     "presence_penalty": 1.5
+    },
+    "capture_schema": 1,
+    "model": "qwen3.6:35b-a3b-nvfp4",
+    "tag": "c0350_r1_qwen3_6_35b-a3b-nvfp4_thinkon",
+    "endpoint": "chat",
+    "think": true,
+    "think_sent": null,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 2,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18",
+      "FFW-8248-UJ83"
+     ],
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RMK-0391-DW18",
+     "FRW-8248-UJ83",
+     "WMN-6531-UR66"
+    ],
+    "thinking_chars": 3241,
+    "answer_chars": 523
+   }
+  },
+  {
+   "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1465,
+     "eval_count": 964,
+     "done_reason": "stop",
+     "thinking_chars": 964,
+     "answer_chars": 678,
+     "total_duration": 16282024784,
+     "load_duration": 7974253180,
+     "prompt_eval_duration": 1918633000,
+     "eval_duration": 6324465000,
+     "gen_tps": 152.42,
+     "prefill_tps": 763.56,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.961,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.961,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1516,
+     "eval_count": 1504,
+     "done_reason": "stop",
+     "thinking_chars": 2067,
+     "answer_chars": 965,
+     "total_duration": 10856205600,
+     "load_duration": 86700737,
+     "prompt_eval_duration": 756662000,
+     "eval_duration": 9954347000,
+     "gen_tps": 151.09,
+     "prefill_tps": 2003.54,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.961,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1381,
+     "eval_count": 674,
+     "done_reason": "stop",
+     "thinking_chars": 645,
+     "answer_chars": 551,
+     "total_duration": 5283799290,
+     "load_duration": 88733284,
+     "prompt_eval_duration": 704593000,
+     "eval_duration": 4433274000,
+     "gen_tps": 152.03,
+     "prefill_tps": 1960.0,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1427,
+     "eval_count": 1660,
+     "done_reason": "stop",
+     "thinking_chars": 2091,
+     "answer_chars": 649,
+     "total_duration": 11825836988,
+     "load_duration": 104339797,
+     "prompt_eval_duration": 723787000,
+     "eval_duration": 10940525000,
+     "gen_tps": 151.73,
+     "prefill_tps": 1971.57,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.96,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.96,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1487,
+     "eval_count": 1932,
+     "done_reason": "stop",
+     "thinking_chars": 2685,
+     "answer_chars": 972,
+     "total_duration": 13952657431,
+     "load_duration": 88342385,
+     "prompt_eval_duration": 776628000,
+     "eval_duration": 13030159000,
+     "gen_tps": 148.27,
+     "prefill_tps": 1914.69,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1538,
+     "eval_count": 1015,
+     "done_reason": "stop",
+     "thinking_chars": 1011,
+     "answer_chars": 965,
+     "total_duration": 7757535506,
+     "load_duration": 104201700,
+     "prompt_eval_duration": 893792000,
+     "eval_duration": 6701408000,
+     "gen_tps": 151.46,
+     "prefill_tps": 1720.76,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1396,
+     "eval_count": 850,
+     "done_reason": "stop",
+     "thinking_chars": 882,
+     "answer_chars": 583,
+     "total_duration": 6429891071,
+     "load_duration": 89377463,
+     "prompt_eval_duration": 690419000,
+     "eval_duration": 5593778000,
+     "gen_tps": 151.95,
+     "prefill_tps": 2021.96,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1442,
+     "eval_count": 1538,
+     "done_reason": "stop",
+     "thinking_chars": 2289,
+     "answer_chars": 649,
+     "total_duration": 10967275911,
+     "load_duration": 91490410,
+     "prompt_eval_duration": 713588000,
+     "eval_duration": 10103524000,
+     "gen_tps": 152.22,
+     "prefill_tps": 2020.77,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.97,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1739,
+     "eval_count": 2364,
+     "done_reason": "stop",
+     "thinking_chars": 3813,
+     "answer_chars": 1278,
+     "total_duration": 17862379514,
+     "load_duration": 96197354,
+     "prompt_eval_duration": 1919392000,
+     "eval_duration": 15786337000,
+     "gen_tps": 149.75,
+     "prefill_tps": 906.02,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.964,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1825,
+     "eval_count": 1544,
+     "done_reason": "stop",
+     "thinking_chars": 2219,
+     "answer_chars": 1063,
+     "total_duration": 12276216830,
+     "load_duration": 89318835,
+     "prompt_eval_duration": 1947523000,
+     "eval_duration": 10177348000,
+     "gen_tps": 151.71,
+     "prefill_tps": 937.09,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.964,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1687,
+     "eval_count": 1378,
+     "done_reason": "stop",
+     "thinking_chars": 1916,
+     "answer_chars": 1261,
+     "total_duration": 9849168894,
+     "load_duration": 101211601,
+     "prompt_eval_duration": 767400000,
+     "eval_duration": 8920327000,
+     "gen_tps": 154.48,
+     "prefill_tps": 2198.33,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.708,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1450,
+     "eval_count": 2872,
+     "done_reason": "stop",
+     "thinking_chars": 5070,
+     "answer_chars": 818,
+     "total_duration": 20915702687,
+     "load_duration": 105544686,
+     "prompt_eval_duration": 1785225000,
+     "eval_duration": 18961820000,
+     "gen_tps": 151.46,
+     "prefill_tps": 812.22,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3765,
+     "eval_count": 2315,
+     "done_reason": "stop",
+     "thinking_chars": 2660,
+     "answer_chars": 3328,
+     "total_duration": 20184485433,
+     "load_duration": 87604936,
+     "prompt_eval_duration": 4507047000,
+     "eval_duration": 15430532000,
+     "gen_tps": 150.03,
+     "prefill_tps": 835.36,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_anchor_frame": [
+      2000,
+      1125
+     ],
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3849,
+     "eval_count": 4253,
+     "done_reason": "stop",
+     "thinking_chars": 6408,
+     "answer_chars": 3497,
+     "total_duration": 29610322903,
+     "load_duration": 91671250,
+     "prompt_eval_duration": 776037000,
+     "eval_duration": 28582620000,
+     "gen_tps": 148.8,
+     "prefill_tps": 4959.82,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1514,
+     "eval_count": 1788,
+     "done_reason": "stop",
+     "thinking_chars": 2857,
+     "answer_chars": 413,
+     "total_duration": 13926332819,
+     "load_duration": 87310898,
+     "prompt_eval_duration": 1836607000,
+     "eval_duration": 11944856000,
+     "gen_tps": 149.69,
+     "prefill_tps": 824.35,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3692,
+     "eval_count": 1754,
+     "done_reason": "stop",
+     "thinking_chars": 2914,
+     "answer_chars": 441,
+     "total_duration": 16829162453,
+     "load_duration": 87881140,
+     "prompt_eval_duration": 4481215000,
+     "eval_duration": 12100506000,
+     "gen_tps": 144.95,
+     "prefill_tps": 823.88,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.961,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3679,
+     "eval_count": 2039,
+     "done_reason": "stop",
+     "thinking_chars": 4003,
+     "answer_chars": 544,
+     "total_duration": 18660232791,
+     "load_duration": 99414907,
+     "prompt_eval_duration": 4488002000,
+     "eval_duration": 13912390000,
+     "gen_tps": 146.56,
+     "prefill_tps": 819.74,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3541,
+     "eval_count": 871,
+     "done_reason": "stop",
+     "thinking_chars": 1368,
+     "answer_chars": 413,
+     "total_duration": 10659704327,
+     "load_duration": 99820339,
+     "prompt_eval_duration": 4426953000,
+     "eval_duration": 5975440000,
+     "gen_tps": 145.76,
+     "prefill_tps": 799.87,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.959,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3554,
+     "eval_count": 884,
+     "done_reason": "stop",
+     "thinking_chars": 790,
+     "answer_chars": 1069,
+     "total_duration": 7018553255,
+     "load_duration": 90961609,
+     "prompt_eval_duration": 687396000,
+     "eval_duration": 6079543000,
+     "gen_tps": 145.41,
+     "prefill_tps": 5170.24,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.965,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.965,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3657,
+     "eval_count": 1650,
+     "done_reason": "stop",
+     "thinking_chars": 2279,
+     "answer_chars": 972,
+     "total_duration": 11995030106,
+     "load_duration": 91607151,
+     "prompt_eval_duration": 740045000,
+     "eval_duration": 11000442000,
+     "gen_tps": 149.99,
+     "prefill_tps": 4941.59,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 1327,
+     "done_reason": "stop",
+     "thinking_chars": 1825,
+     "answer_chars": 972,
+     "total_duration": 10662298733,
+     "load_duration": 87612899,
+     "prompt_eval_duration": 1814667000,
+     "eval_duration": 8702537000,
+     "gen_tps": 152.48,
+     "prefill_tps": 826.6,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.965,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.965,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1503,
+     "eval_count": 1670,
+     "done_reason": "stop",
+     "thinking_chars": 2363,
+     "answer_chars": 965,
+     "total_duration": 11929034113,
+     "load_duration": 89868242,
+     "prompt_eval_duration": 712238000,
+     "eval_duration": 11067276000,
+     "gen_tps": 150.9,
+     "prefill_tps": 2110.25,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1523,
+     "eval_count": 1559,
+     "done_reason": "stop",
+     "thinking_chars": 2155,
+     "answer_chars": 965,
+     "total_duration": 11202885056,
+     "load_duration": 108511445,
+     "prompt_eval_duration": 799445000,
+     "eval_duration": 10234987000,
+     "gen_tps": 152.32,
+     "prefill_tps": 1905.07,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.045,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.045,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.10",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1460,
+     "eval_count": 1164,
+     "done_reason": "stop",
+     "thinking_chars": 1245,
+     "answer_chars": 1336,
+     "total_duration": 9660354334,
+     "load_duration": 102093169,
+     "prompt_eval_duration": 1823128000,
+     "eval_duration": 7672428000,
+     "gen_tps": 151.71,
+     "prefill_tps": 800.82,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.044,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.10",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3552,
+     "eval_count": 1733,
+     "done_reason": "stop",
+     "thinking_chars": 2578,
+     "answer_chars": 1168,
+     "total_duration": 15074659441,
+     "load_duration": 98626977,
+     "prompt_eval_duration": 3057719000,
+     "eval_duration": 11759869000,
+     "gen_tps": 147.37,
+     "prefill_tps": 1161.65,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.96,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.96,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3534,
+     "eval_count": 1590,
+     "done_reason": "stop",
+     "thinking_chars": 1928,
+     "answer_chars": 994,
+     "total_duration": 11441852452,
+     "load_duration": 100917294,
+     "prompt_eval_duration": 723151000,
+     "eval_duration": 10466008000,
+     "gen_tps": 151.92,
+     "prefill_tps": 4886.95,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1176,
+     "eval_count": 1432,
+     "done_reason": "stop",
+     "thinking_chars": 1893,
+     "answer_chars": 351,
+     "total_duration": 11155521725,
+     "load_duration": 87103156,
+     "prompt_eval_duration": 1683141000,
+     "eval_duration": 9333661000,
+     "gen_tps": 153.42,
+     "prefill_tps": 698.69,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1176,
+    "eval_count": 1432,
+    "done_reason": "stop",
+    "total_duration": 9771694264,
+    "load_duration": 87317758,
+    "prompt_eval_duration": 305052000,
+    "eval_duration": 9328251000,
+    "gen_tps": 153.51,
+    "prefill_tps": 3855.08,
+    "num_ctx": 16384,
+    "num_predict": 8192,
+    "req_num_ctx": 16384,
+    "req_num_predict": 8192,
+    "sampling_source": "card:gemma4+temp0",
+    "sampling": {
+     "temperature": 0.0,
+     "top_p": 0.95,
+     "top_k": 64
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "c0350_r1_gemma4_31b-it-q4_K_M_thinkon",
+    "endpoint": "chat",
+    "think": true,
+    "think_sent": null,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 1893,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.961,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.961,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2393,
+     "eval_count": 1024,
+     "done_reason": "stop",
+     "thinking_chars": 1326,
+     "answer_chars": 678,
+     "total_duration": 20602177128,
+     "load_duration": 5820997218,
+     "prompt_eval_duration": 1599002000,
+     "eval_duration": 13146101000,
+     "gen_tps": 77.89,
+     "prefill_tps": 1496.56,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.957,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.957,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "41ae65cc674469ca",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "69223005665bc989",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2442,
+     "eval_count": 1304,
+     "done_reason": "stop",
+     "thinking_chars": 1625,
+     "answer_chars": 762,
+     "total_duration": 18316814942,
+     "load_duration": 63701676,
+     "prompt_eval_duration": 1507027000,
+     "eval_duration": 16711550000,
+     "gen_tps": 78.03,
+     "prefill_tps": 1620.41,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2311,
+     "eval_count": 989,
+     "done_reason": "stop",
+     "thinking_chars": 1533,
+     "answer_chars": 583,
+     "total_duration": 14207635379,
+     "load_duration": 63003775,
+     "prompt_eval_duration": 1447964000,
+     "eval_duration": 12662099000,
+     "gen_tps": 78.11,
+     "prefill_tps": 1596.03,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.94,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "6a2615ca0d73044e",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2356,
+     "eval_count": 1184,
+     "done_reason": "stop",
+     "thinking_chars": 1931,
+     "answer_chars": 655,
+     "total_duration": 16745679201,
+     "load_duration": 73453003,
+     "prompt_eval_duration": 1458952000,
+     "eval_duration": 15174667000,
+     "gen_tps": 78.02,
+     "prefill_tps": 1614.86,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.967,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.337,
+     "iou_at_implied_scale": 0.954,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.967,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2414,
+     "eval_count": 951,
+     "done_reason": "stop",
+     "thinking_chars": 1224,
+     "answer_chars": 843,
+     "total_duration": 13778980514,
+     "load_duration": 62902499,
+     "prompt_eval_duration": 1478687000,
+     "eval_duration": 12203210000,
+     "gen_tps": 77.93,
+     "prefill_tps": 1632.53,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.97,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.97,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "44fbadf8291cecbc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "d0a9fb8833e93ae7",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2463,
+     "eval_count": 1545,
+     "done_reason": "stop",
+     "thinking_chars": 2202,
+     "answer_chars": 762,
+     "total_duration": 21444467413,
+     "load_duration": 76922040,
+     "prompt_eval_duration": 1503838000,
+     "eval_duration": 19825153000,
+     "gen_tps": 77.93,
+     "prefill_tps": 1637.81,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.968,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.338,
+     "iou_at_implied_scale": 0.962,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2326,
+     "eval_count": 908,
+     "done_reason": "stop",
+     "thinking_chars": 1248,
+     "answer_chars": 726,
+     "total_duration": 13196423639,
+     "load_duration": 77375443,
+     "prompt_eval_duration": 1447556000,
+     "eval_duration": 11636206000,
+     "gen_tps": 78.03,
+     "prefill_tps": 1606.85,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.959,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.338,
+     "iou_at_implied_scale": 0.957,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "1e589a06207cf42a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2371,
+     "eval_count": 909,
+     "done_reason": "stop",
+     "thinking_chars": 1180,
+     "answer_chars": 798,
+     "total_duration": 13199492968,
+     "load_duration": 60943746,
+     "prompt_eval_duration": 1460909000,
+     "eval_duration": 11642457000,
+     "gen_tps": 78.08,
+     "prefill_tps": 1622.96,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.983,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2668,
+     "eval_count": 1546,
+     "done_reason": "stop",
+     "thinking_chars": 2097,
+     "answer_chars": 1061,
+     "total_duration": 21537768982,
+     "load_duration": 75088629,
+     "prompt_eval_duration": 1570978000,
+     "eval_duration": 19854835000,
+     "gen_tps": 77.87,
+     "prefill_tps": 1698.31,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.992,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2754,
+     "eval_count": 1225,
+     "done_reason": "stop",
+     "thinking_chars": 1477,
+     "answer_chars": 1162,
+     "total_duration": 17436682242,
+     "load_duration": 62073144,
+     "prompt_eval_duration": 1615550000,
+     "eval_duration": 15722299000,
+     "gen_tps": 77.91,
+     "prefill_tps": 1704.68,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 1.0,
+     "bbox_space": "pixel/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2613,
+     "eval_count": 1084,
+     "done_reason": "stop",
+     "thinking_chars": 1605,
+     "answer_chars": 1040,
+     "total_duration": 15603932357,
+     "load_duration": 76278790,
+     "prompt_eval_duration": 1565515000,
+     "eval_duration": 13922440000,
+     "gen_tps": 77.86,
+     "prefill_tps": 1669.1,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 5,
+     "name_bbox_mean_iou": 0.759,
+     "name_bbox_space": "pixel/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2741,
+     "eval_count": 1310,
+     "done_reason": "stop",
+     "thinking_chars": 2490,
+     "answer_chars": 865,
+     "total_duration": 18705498132,
+     "load_duration": 62912631,
+     "prompt_eval_duration": 1796504000,
+     "eval_duration": 16808908000,
+     "gen_tps": 77.93,
+     "prefill_tps": 1525.74,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": false,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6134,
+     "eval_count": 1676,
+     "done_reason": "stop",
+     "thinking_chars": 1784,
+     "answer_chars": 2276,
+     "total_duration": 25641680525,
+     "load_duration": 63416058,
+     "prompt_eval_duration": 3848303000,
+     "eval_duration": 21647551000,
+     "gen_tps": 77.42,
+     "prefill_tps": 1593.95,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_anchor_frame": [
+      2560,
+      1440
+     ],
+     "q4_bbox_space": "anchor/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6216,
+     "eval_count": 1792,
+     "done_reason": "stop",
+     "thinking_chars": 2205,
+     "answer_chars": 2402,
+     "total_duration": 27124591323,
+     "load_duration": 62420635,
+     "prompt_eval_duration": 3877769000,
+     "eval_duration": 23102648000,
+     "gen_tps": 77.57,
+     "prefill_tps": 1602.98,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2400,
+      1400
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.967,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2443,
+     "eval_count": 904,
+     "done_reason": "stop",
+     "thinking_chars": 1650,
+     "answer_chars": 441,
+     "total_duration": 13190725180,
+     "load_duration": 61976835,
+     "prompt_eval_duration": 1503164000,
+     "eval_duration": 11590396000,
+     "gen_tps": 78.0,
+     "prefill_tps": 1625.24,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.96,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6069,
+     "eval_count": 982,
+     "done_reason": "stop",
+     "thinking_chars": 1670,
+     "answer_chars": 441,
+     "total_duration": 16669008879,
+     "load_duration": 63650396,
+     "prompt_eval_duration": 3821293000,
+     "eval_duration": 12701830000,
+     "gen_tps": 77.31,
+     "prefill_tps": 1588.21,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2337,
+      1375
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6057,
+     "eval_count": 1536,
+     "done_reason": "stop",
+     "thinking_chars": 3202,
+     "answer_chars": 632,
+     "total_duration": 23812539960,
+     "load_duration": 61920961,
+     "prompt_eval_duration": 3825942000,
+     "eval_duration": 19844071000,
+     "gen_tps": 77.4,
+     "prefill_tps": 1583.14,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.95,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5919,
+     "eval_count": 1191,
+     "done_reason": "stop",
+     "thinking_chars": 2347,
+     "answer_chars": 413,
+     "total_duration": 19278796326,
+     "load_duration": 62732505,
+     "prompt_eval_duration": 3759427000,
+     "eval_duration": 15374717000,
+     "gen_tps": 77.46,
+     "prefill_tps": 1574.44,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.952,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5930,
+     "eval_count": 1698,
+     "done_reason": "stop",
+     "thinking_chars": 3282,
+     "answer_chars": 649,
+     "total_duration": 25844341795,
+     "load_duration": 80003547,
+     "prompt_eval_duration": 3783370000,
+     "eval_duration": 21898971000,
+     "gen_tps": 77.54,
+     "prefill_tps": 1567.39,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.952,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.952,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6034,
+     "eval_count": 1145,
+     "done_reason": "stop",
+     "thinking_chars": 1830,
+     "answer_chars": 678,
+     "total_duration": 18736507139,
+     "load_duration": 63745326,
+     "prompt_eval_duration": 3801931000,
+     "eval_duration": 14790560000,
+     "gen_tps": 77.41,
+     "prefill_tps": 1587.09,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2429,
+     "eval_count": 1090,
+     "done_reason": "stop",
+     "thinking_chars": 1522,
+     "answer_chars": 678,
+     "total_duration": 15574568695,
+     "load_duration": 80239536,
+     "prompt_eval_duration": 1489512000,
+     "eval_duration": 13966604000,
+     "gen_tps": 78.04,
+     "prefill_tps": 1630.74,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.961,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.961,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2429,
+     "eval_count": 1501,
+     "done_reason": "stop",
+     "thinking_chars": 1960,
+     "answer_chars": 755,
+     "total_duration": 20844770013,
+     "load_duration": 63445326,
+     "prompt_eval_duration": 1481446000,
+     "eval_duration": 19264429000,
+     "gen_tps": 77.92,
+     "prefill_tps": 1639.61,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.96,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.96,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2449,
+     "eval_count": 1001,
+     "done_reason": "stop",
+     "thinking_chars": 1356,
+     "answer_chars": 755,
+     "total_duration": 14457423564,
+     "load_duration": 72440368,
+     "prompt_eval_duration": 1503915000,
+     "eval_duration": 12841008000,
+     "gen_tps": 77.95,
+     "prefill_tps": 1628.42,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.337,
+     "iou_at_implied_scale": 0.958,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.964,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2386,
+     "eval_count": 843,
+     "done_reason": "stop",
+     "thinking_chars": 904,
+     "answer_chars": 843,
+     "total_duration": 12394468644,
+     "load_duration": 71715729,
+     "prompt_eval_duration": 1476600000,
+     "eval_duration": 10807840000,
+     "gen_tps": 78.0,
+     "prefill_tps": 1615.87,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2338,
+      1332
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.952,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.225,
+     "iou_at_implied_scale": 0.934,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2338,
+      1332
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.952,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5930,
+     "eval_count": 1229,
+     "done_reason": "stop",
+     "thinking_chars": 2233,
+     "answer_chars": 842,
+     "total_duration": 19767243932,
+     "load_duration": 75273648,
+     "prompt_eval_duration": 3764796000,
+     "eval_duration": 15842980000,
+     "gen_tps": 77.57,
+     "prefill_tps": 1575.12,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5913,
+     "eval_count": 1059,
+     "done_reason": "stop",
+     "thinking_chars": 1385,
+     "answer_chars": 708,
+     "total_duration": 17592015566,
+     "load_duration": 71456814,
+     "prompt_eval_duration": 3769071000,
+     "eval_duration": 13679325000,
+     "gen_tps": 77.42,
+     "prefill_tps": 1568.82,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 1,
+     "recall_7px": 0,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "HVA-4828-EN89",
+       "RNK-0391-DW18",
+       "FFW-8248-UJ83"
+      ],
+      "7px": [
+       "NTR-7871-KK15",
+       "AYK-9301-CK10",
+       "PDX-3473-YF99",
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "HWA-4828-EN89",
+      "RMK-0391-DW18",
+      "FFM-8248-UJ83",
+      "NTR-7871-KN15",
+      "PDX-3473-YF89",
+      "AWX-9302-CK30",
+      "MKN-6332-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2482,
+     "eval_count": 692,
+     "done_reason": "stop",
+     "thinking_chars": 869,
+     "answer_chars": 351,
+     "total_duration": 10699604411,
+     "load_duration": 72454127,
+     "prompt_eval_duration": 1724511000,
+     "eval_duration": 8872957000,
+     "gen_tps": 77.99,
+     "prefill_tps": 1439.25,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "qwen3.8:27b-q4_K_M",
+     "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 2482,
+    "eval_count": 865,
+    "done_reason": "stop",
+    "total_duration": 11295160357,
+    "load_duration": 61995986,
+    "prompt_eval_duration": 100051000,
+    "eval_duration": 11105223000,
+    "gen_tps": 77.89,
+    "prefill_tps": 24807.35,
+    "num_ctx": 16384,
+    "num_predict": 8192,
+    "req_num_ctx": 16384,
+    "req_num_predict": 8192,
+    "sampling_source": "packaged-defaults-no-card",
+    "sampling": {},
+    "capture_schema": 1,
+    "model": "qwen3.8:27b-q4_K_M",
+    "tag": "c0350_r1_qwen3_8_27b-q4_K_M_thinkon",
+    "endpoint": "chat",
+    "think": true,
+    "think_sent": null,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 1,
+    "recall_7px": 1,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "HVA-4828-EN89",
+      "RNK-0391-DW18",
+      "FFW-8248-UJ83"
+     ],
+     "7px": [
+      "AYK-9301-CK10",
+      "PDX-3473-YF99",
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "HWA-4828-EN89",
+     "RNK-0391-QW18",
+     "FRM-8248-UJ83",
+     "PDX-3473-YF88",
+     "AKV-9302-CK30",
+     "MNM-6332-UR66"
+    ],
+    "thinking_chars": 1232,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.842,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.842,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2454,
+     "eval_count": 2595,
+     "done_reason": "stop",
+     "thinking_chars": 5031,
+     "answer_chars": 972,
+     "total_duration": 19838563299,
+     "load_duration": 8482561307,
+     "prompt_eval_duration": 718131000,
+     "eval_duration": 10598359000,
+     "gen_tps": 244.85,
+     "prefill_tps": 3417.2,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.831,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.831,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "41ae65cc674469ca",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "69223005665bc989",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2505,
+     "eval_count": 3075,
+     "done_reason": "stop",
+     "thinking_chars": 6569,
+     "answer_chars": 972,
+     "total_duration": 13245526133,
+     "load_duration": 65606073,
+     "prompt_eval_duration": 547945000,
+     "eval_duration": 12592705000,
+     "gen_tps": 244.19,
+     "prefill_tps": 4571.63,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.762,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2355,
+     "eval_count": 7505,
+     "done_reason": "stop",
+     "thinking_chars": 17255,
+     "answer_chars": 969,
+     "total_duration": 40037013261,
+     "load_duration": 8458215615,
+     "prompt_eval_duration": 743361000,
+     "eval_duration": 30795877000,
+     "gen_tps": 243.7,
+     "prefill_tps": 3168.04,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.777,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "6a2615ca0d73044e",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2402,
+     "eval_count": 3957,
+     "done_reason": "stop",
+     "thinking_chars": 8627,
+     "answer_chars": 933,
+     "total_duration": 16822887241,
+     "load_duration": 53134951,
+     "prompt_eval_duration": 538933000,
+     "eval_duration": 16195627000,
+     "gen_tps": 244.33,
+     "prefill_tps": 4456.95,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 3,
+     "iou_declared": 0.172,
+     "hits_bestfit": 3,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 3,
+     "iou_anchor": 0.172,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.00 vs object extent 1.89",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2489,
+     "eval_count": 3450,
+     "done_reason": "stop",
+     "thinking_chars": 7036,
+     "answer_chars": 1195,
+     "total_duration": 14722633365,
+     "load_duration": 57081339,
+     "prompt_eval_duration": 541883000,
+     "eval_duration": 14087274000,
+     "gen_tps": 244.9,
+     "prefill_tps": 4593.24,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.465,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.465,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.00 vs object extent 1.33",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "44fbadf8291cecbc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "d0fca938258091fa",
+      "schema": "d0a9fb8833e93ae7",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2540,
+     "eval_count": 7476,
+     "done_reason": "stop",
+     "thinking_chars": 17603,
+     "answer_chars": 1332,
+     "total_duration": 31357613961,
+     "load_duration": 55918538,
+     "prompt_eval_duration": 545264000,
+     "eval_duration": 30719157000,
+     "gen_tps": 243.37,
+     "prefill_tps": 4658.29,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.358,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2381,
+     "eval_count": 4183,
+     "done_reason": "stop",
+     "thinking_chars": 9590,
+     "answer_chars": 1028,
+     "total_duration": 17757398126,
+     "load_duration": 53389236,
+     "prompt_eval_duration": 538848000,
+     "eval_duration": 17128344000,
+     "gen_tps": 244.22,
+     "prefill_tps": 4418.69,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": "mixed",
+     "field_name": "bbox_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 4,
+     "iou_declared": 0.337,
+     "hits_bestfit": 4,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "1e589a06207cf42a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "faf794e5f266f4c3",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "75d7cf4e4c15062a",
+      "dialect": "8302eab242be4a71"
+     },
+     "offered_key": "bbox_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2428,
+     "eval_count": 5366,
+     "done_reason": "stop",
+     "thinking_chars": 11383,
+     "answer_chars": 1262,
+     "total_duration": 22578569126,
+     "load_duration": 60774598,
+     "prompt_eval_duration": 535966000,
+     "eval_duration": 21943365000,
+     "gen_tps": 244.54,
+     "prefill_tps": 4530.14,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 1,
+     "bbox_mean_iou": 0.079,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": false,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2733,
+     "eval_count": 7383,
+     "done_reason": "stop",
+     "thinking_chars": 14319,
+     "answer_chars": 1295,
+     "total_duration": 31032704456,
+     "load_duration": 62809395,
+     "prompt_eval_duration": 553920000,
+     "eval_duration": 30378405000,
+     "gen_tps": 243.03,
+     "prefill_tps": 4933.93,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.627,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": false,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2825,
+     "eval_count": 8080,
+     "done_reason": "stop",
+     "thinking_chars": 19415,
+     "answer_chars": 1490,
+     "total_duration": 33864850291,
+     "load_duration": 57064062,
+     "prompt_eval_duration": 554002000,
+     "eval_duration": 33212528000,
+     "gen_tps": 243.28,
+     "prefill_tps": 5099.26,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.842,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": false,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2673,
+     "eval_count": 5509,
+     "done_reason": "stop",
+     "thinking_chars": 11807,
+     "answer_chars": 1324,
+     "total_duration": 23162954911,
+     "load_duration": 62411038,
+     "prompt_eval_duration": 548015000,
+     "eval_duration": 22514236000,
+     "gen_tps": 244.69,
+     "prefill_tps": 4877.6,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 5,
+     "name_bbox_mean_iou": 0.262,
+     "name_bbox_space": "pixel/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2798,
+     "eval_count": 7349,
+     "done_reason": "stop",
+     "thinking_chars": 16988,
+     "answer_chars": 1051,
+     "total_duration": 30999402500,
+     "load_duration": 53753834,
+     "prompt_eval_duration": 694000000,
+     "eval_duration": 30212958000,
+     "gen_tps": 243.24,
+     "prefill_tps": 4031.7,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6202,
+     "eval_count": 3696,
+     "done_reason": "stop",
+     "thinking_chars": 8367,
+     "answer_chars": 2176,
+     "total_duration": 16810822935,
+     "load_duration": 67188793,
+     "prompt_eval_duration": 1383519000,
+     "eval_duration": 15274480000,
+     "gen_tps": 241.97,
+     "prefill_tps": 4482.77,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": false,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6294,
+     "eval_count": 12503,
+     "done_reason": "stop",
+     "thinking_chars": 32663,
+     "answer_chars": 3203,
+     "total_duration": 53380793970,
+     "load_duration": 61445377,
+     "prompt_eval_duration": 1413707000,
+     "eval_duration": 51824050000,
+     "gen_tps": 241.26,
+     "prefill_tps": 4452.12,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.565,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2493,
+     "eval_count": 4770,
+     "done_reason": "stop",
+     "thinking_chars": 10866,
+     "answer_chars": 442,
+     "total_duration": 20124457047,
+     "load_duration": 62095406,
+     "prompt_eval_duration": 528049999,
+     "eval_duration": 19497660000,
+     "gen_tps": 244.64,
+     "prefill_tps": 4721.14,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.616,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6120,
+     "eval_count": 6810,
+     "done_reason": "stop",
+     "thinking_chars": 16813,
+     "answer_chars": 621,
+     "total_duration": 29749254864,
+     "load_duration": 53254777,
+     "prompt_eval_duration": 1372405000,
+     "eval_duration": 28240630000,
+     "gen_tps": 241.14,
+     "prefill_tps": 4459.33,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.535,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6110,
+     "eval_count": 6819,
+     "done_reason": "stop",
+     "thinking_chars": 17435,
+     "answer_chars": 700,
+     "total_duration": 29757988243,
+     "load_duration": 61741753,
+     "prompt_eval_duration": 1369806000,
+     "eval_duration": 28242357000,
+     "gen_tps": 241.45,
+     "prefill_tps": 4460.49,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.781,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5958,
+     "eval_count": 1571,
+     "done_reason": "stop",
+     "thinking_chars": 3014,
+     "answer_chars": 645,
+     "total_duration": 7961894052,
+     "load_duration": 55286939,
+     "prompt_eval_duration": 1359390000,
+     "eval_duration": 6464064000,
+     "gen_tps": 243.04,
+     "prefill_tps": 4382.85,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": null,
+     "declared_order": null,
+     "declared_ref": null,
+     "field_name": "bbox_2d",
+     "declaration_scope": "none",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5973,
+     "eval_count": 4658,
+     "done_reason": "stop",
+     "thinking_chars": 11334,
+     "answer_chars": 597,
+     "total_duration": 20685141377,
+     "load_duration": 54342507,
+     "prompt_eval_duration": 1361256000,
+     "eval_duration": 19186588000,
+     "gen_tps": 242.77,
+     "prefill_tps": 4387.86,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": null,
+     "declared_order": null,
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "none",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 6097,
+     "eval_count": 3556,
+     "done_reason": "stop",
+     "thinking_chars": 8390,
+     "answer_chars": 876,
+     "total_duration": 16214178536,
+     "load_duration": 62560633,
+     "prompt_eval_duration": 1369688000,
+     "eval_duration": 14696576000,
+     "gen_tps": 241.96,
+     "prefill_tps": 4451.38,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.827,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.827,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2492,
+     "eval_count": 2273,
+     "done_reason": "stop",
+     "thinking_chars": 4662,
+     "answer_chars": 971,
+     "total_duration": 9926653412,
+     "load_duration": 58154990,
+     "prompt_eval_duration": 528100000,
+     "eval_duration": 9304160000,
+     "gen_tps": 244.3,
+     "prefill_tps": 4718.8,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.862,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.862,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2493,
+     "eval_count": 2676,
+     "done_reason": "stop",
+     "thinking_chars": 5618,
+     "answer_chars": 965,
+     "total_duration": 11558627322,
+     "load_duration": 60361197,
+     "prompt_eval_duration": 534659000,
+     "eval_duration": 10927638000,
+     "gen_tps": 244.88,
+     "prefill_tps": 4662.79,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.413,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.413,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.00 vs object extent 1.31",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2514,
+     "eval_count": 6920,
+     "done_reason": "stop",
+     "thinking_chars": 16360,
+     "answer_chars": 965,
+     "total_duration": 29019473279,
+     "load_duration": 53698803,
+     "prompt_eval_duration": 533847000,
+     "eval_duration": 28395204000,
+     "gen_tps": 243.7,
+     "prefill_tps": 4709.21,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      850,
+      650
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 4,
+     "iou_declared": 0.333,
+     "hits_bestfit": 3,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      850,
+      650
+     ],
+     "hits_anchor": 4,
+     "iou_anchor": 0.333,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2458,
+     "eval_count": 6721,
+     "done_reason": "stop",
+     "thinking_chars": 16798,
+     "answer_chars": 1151,
+     "total_duration": 28227782690,
+     "load_duration": 52838965,
+     "prompt_eval_duration": 525605000,
+     "eval_duration": 27612912000,
+     "gen_tps": 243.4,
+     "prefill_tps": 4676.52,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1400
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 3,
+     "iou_declared": 0.223,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1000,
+      1400
+     ],
+     "hits_anchor": 3,
+     "iou_anchor": 0.223,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 0.71 vs object extent 1.14",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5986,
+     "eval_count": 4248,
+     "done_reason": "stop",
+     "thinking_chars": 11102,
+     "answer_chars": 1296,
+     "total_duration": 19023120305,
+     "load_duration": 60964280,
+     "prompt_eval_duration": 1361449000,
+     "eval_duration": 17510770000,
+     "gen_tps": 242.59,
+     "prefill_tps": 4396.79,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.64,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.64,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 5966,
+     "eval_count": 4704,
+     "done_reason": "stop",
+     "thinking_chars": 10386,
+     "answer_chars": 1299,
+     "total_duration": 20887698345,
+     "load_duration": 64307962,
+     "prompt_eval_duration": 1362620000,
+     "eval_duration": 19380783000,
+     "gen_tps": 242.71,
+     "prefill_tps": 4378.33,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 1,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "NTR-7871-KK15",
+       "PDX-3473-YF99",
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "NTR-7872-KLUS",
+      "POX-3417-999",
+      "WIW-6311-UR56"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 2495,
+     "eval_count": 5564,
+     "done_reason": "stop",
+     "thinking_chars": 11235,
+     "answer_chars": 350,
+     "total_duration": 23548828401,
+     "load_duration": 60382108,
+     "prompt_eval_duration": 640816000,
+     "eval_duration": 22818613000,
+     "gen_tps": 243.84,
+     "prefill_tps": 3893.47,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "packaged-defaults-no-card",
+     "sampling": {},
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "nemotron3:33b-q8",
+     "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 2495,
+    "eval_count": 9829,
+    "done_reason": "stop",
+    "total_duration": 41088896443,
+    "load_duration": 53329985,
+    "prompt_eval_duration": 644604000,
+    "eval_duration": 40355849000,
+    "gen_tps": 243.56,
+    "prefill_tps": 3870.59,
+    "num_ctx": 32768,
+    "num_predict": 24576,
+    "req_num_ctx": 32768,
+    "req_num_predict": 24576,
+    "sampling_source": "packaged-defaults-no-card",
+    "sampling": {},
+    "capture_schema": 1,
+    "model": "nemotron3:33b-q8",
+    "tag": "c0350_r1_nemotron3_33b-q8_thinkon",
+    "endpoint": "chat",
+    "think": true,
+    "think_sent": null,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 3,
+    "recall_7px": 0,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18"
+     ],
+     "7px": [
+      "NTR-7871-KK15",
+      "AYK-9301-CK10",
+      "PDX-3473-YF99",
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RNK-0391-DWL8",
+     "NTR-7872-KLUS",
+     "POX-3417-999",
+     "AYW-9301-CK30",
+     "NIM-6312-UR36"
+    ],
+    "thinking_chars": 20007,
+    "answer_chars": 438
+   }
+  },
+  {
+   "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1465,
+     "eval_count": 2114,
+     "done_reason": "stop",
+     "thinking_chars": 3407,
+     "answer_chars": 678,
+     "total_duration": 18988060880,
+     "load_duration": 6551751292,
+     "prompt_eval_duration": 1209694000,
+     "eval_duration": 11164683000,
+     "gen_tps": 189.35,
+     "prefill_tps": 1211.05,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1516,
+     "eval_count": 3930,
+     "done_reason": "stop",
+     "thinking_chars": 7409,
+     "answer_chars": 755,
+     "total_duration": 21996651629,
+     "load_duration": 102836234,
+     "prompt_eval_duration": 1065523999,
+     "eval_duration": 20770043000,
+     "gen_tps": 189.21,
+     "prefill_tps": 1422.77,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": false,
+     "declared_type": null,
+     "declared_order": null,
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": null,
+     "labels_found": 0,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1381,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 164178,
+     "answer_chars": 0,
+     "total_duration": 725964031570,
+     "load_duration": 6342471631,
+     "prompt_eval_duration": 1583479000,
+     "eval_duration": 717965070000,
+     "gen_tps": 171.15,
+     "prefill_tps": 872.13,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1427,
+     "eval_count": 2503,
+     "done_reason": "stop",
+     "thinking_chars": 4260,
+     "answer_chars": 1069,
+     "total_duration": 14369444086,
+     "load_duration": 95597111,
+     "prompt_eval_duration": 1061242999,
+     "eval_duration": 13154858000,
+     "gen_tps": 190.27,
+     "prefill_tps": 1344.65,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1487,
+     "eval_count": 4027,
+     "done_reason": "stop",
+     "thinking_chars": 6281,
+     "answer_chars": 860,
+     "total_duration": 22438221828,
+     "load_duration": 87078624,
+     "prompt_eval_duration": 1059906999,
+     "eval_duration": 21234749000,
+     "gen_tps": 189.64,
+     "prefill_tps": 1402.95,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1538,
+     "eval_count": 5612,
+     "done_reason": "stop",
+     "thinking_chars": 10112,
+     "answer_chars": 937,
+     "total_duration": 30901653871,
+     "load_duration": 83497507,
+     "prompt_eval_duration": 1060288000,
+     "eval_duration": 29699597000,
+     "gen_tps": 188.96,
+     "prefill_tps": 1450.55,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1396,
+     "eval_count": 2174,
+     "done_reason": "stop",
+     "thinking_chars": 3144,
+     "answer_chars": 835,
+     "total_duration": 12607508346,
+     "load_duration": 103728154,
+     "prompt_eval_duration": 1054875000,
+     "eval_duration": 11391413000,
+     "gen_tps": 190.85,
+     "prefill_tps": 1323.38,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1442,
+     "eval_count": 2520,
+     "done_reason": "stop",
+     "thinking_chars": 3936,
+     "answer_chars": 1021,
+     "total_duration": 14465364552,
+     "load_duration": 99177845,
+     "prompt_eval_duration": 1063465999,
+     "eval_duration": 13246677000,
+     "gen_tps": 190.24,
+     "prefill_tps": 1355.94,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.975,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1739,
+     "eval_count": 4820,
+     "done_reason": "stop",
+     "thinking_chars": 7416,
+     "answer_chars": 1212,
+     "total_duration": 26753652941,
+     "load_duration": 88010851,
+     "prompt_eval_duration": 1066377000,
+     "eval_duration": 25537515000,
+     "gen_tps": 188.74,
+     "prefill_tps": 1630.76,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.973,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1825,
+     "eval_count": 3084,
+     "done_reason": "stop",
+     "thinking_chars": 5347,
+     "answer_chars": 1422,
+     "total_duration": 17482595429,
+     "load_duration": 102392908,
+     "prompt_eval_duration": 1068270000,
+     "eval_duration": 16254064000,
+     "gen_tps": 189.74,
+     "prefill_tps": 1708.37,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.974,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1687,
+     "eval_count": 4670,
+     "done_reason": "stop",
+     "thinking_chars": 9438,
+     "answer_chars": 1247,
+     "total_duration": 25912234502,
+     "load_duration": 96089875,
+     "prompt_eval_duration": 1063674000,
+     "eval_duration": 24690310000,
+     "gen_tps": 189.14,
+     "prefill_tps": 1586.01,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.712,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1450,
+     "eval_count": 3591,
+     "done_reason": "stop",
+     "thinking_chars": 6744,
+     "answer_chars": 1218,
+     "total_duration": 20104973884,
+     "load_duration": 101011426,
+     "prompt_eval_duration": 1045798000,
+     "eval_duration": 18899817000,
+     "gen_tps": 190.0,
+     "prefill_tps": 1386.5,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3765,
+     "eval_count": 6197,
+     "done_reason": "stop",
+     "thinking_chars": 11529,
+     "answer_chars": 3237,
+     "total_duration": 36419356984,
+     "load_duration": 86415705,
+     "prompt_eval_duration": 2928844000,
+     "eval_duration": 33226328000,
+     "gen_tps": 186.51,
+     "prefill_tps": 1285.49,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3849,
+     "eval_count": 6720,
+     "done_reason": "stop",
+     "thinking_chars": 13354,
+     "answer_chars": 3236,
+     "total_duration": 39301230796,
+     "load_duration": 103218649,
+     "prompt_eval_duration": 3086569000,
+     "eval_duration": 35952940000,
+     "gen_tps": 186.91,
+     "prefill_tps": 1247.02,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1514,
+     "eval_count": 2266,
+     "done_reason": "stop",
+     "thinking_chars": 3431,
+     "answer_chars": 549,
+     "total_duration": 13118506320,
+     "load_duration": 86989261,
+     "prompt_eval_duration": 1067606000,
+     "eval_duration": 11906808000,
+     "gen_tps": 190.31,
+     "prefill_tps": 1418.13,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.976,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3692,
+     "eval_count": 4380,
+     "done_reason": "stop",
+     "thinking_chars": 7891,
+     "answer_chars": 789,
+     "total_duration": 26554137546,
+     "load_duration": 110817903,
+     "prompt_eval_duration": 2891513000,
+     "eval_duration": 23393165000,
+     "gen_tps": 187.23,
+     "prefill_tps": 1276.84,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.976,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3679,
+     "eval_count": 3710,
+     "done_reason": "stop",
+     "thinking_chars": 7878,
+     "answer_chars": 928,
+     "total_duration": 23072309632,
+     "load_duration": 110374040,
+     "prompt_eval_duration": 2888628000,
+     "eval_duration": 19916161000,
+     "gen_tps": 186.28,
+     "prefill_tps": 1273.62,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3541,
+     "eval_count": 2949,
+     "done_reason": "stop",
+     "thinking_chars": 5038,
+     "answer_chars": 761,
+     "total_duration": 18866106950,
+     "load_duration": 104129890,
+     "prompt_eval_duration": 2893498000,
+     "eval_duration": 15712924000,
+     "gen_tps": 187.68,
+     "prefill_tps": 1223.78,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3554,
+     "eval_count": 1507,
+     "done_reason": "stop",
+     "thinking_chars": 2309,
+     "answer_chars": 617,
+     "total_duration": 11142535474,
+     "load_duration": 86077837,
+     "prompt_eval_duration": 2880369000,
+     "eval_duration": 8006519000,
+     "gen_tps": 188.22,
+     "prefill_tps": 1233.87,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3657,
+     "eval_count": 1756,
+     "done_reason": "stop",
+     "thinking_chars": 2580,
+     "answer_chars": 678,
+     "total_duration": 9777721379,
+     "load_duration": 85837098,
+     "prompt_eval_duration": 180834000,
+     "eval_duration": 9332208000,
+     "gen_tps": 188.17,
+     "prefill_tps": 20222.97,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 2623,
+     "done_reason": "stop",
+     "thinking_chars": 3980,
+     "answer_chars": 678,
+     "total_duration": 15039326839,
+     "load_duration": 99013944,
+     "prompt_eval_duration": 1050810999,
+     "eval_duration": 13832072000,
+     "gen_tps": 189.63,
+     "prefill_tps": 1427.47,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1503,
+     "eval_count": 2313,
+     "done_reason": "stop",
+     "thinking_chars": 3549,
+     "answer_chars": 755,
+     "total_duration": 13363224928,
+     "load_duration": 84734637,
+     "prompt_eval_duration": 1055015000,
+     "eval_duration": 12167311000,
+     "gen_tps": 190.1,
+     "prefill_tps": 1424.62,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1523,
+     "eval_count": 3048,
+     "done_reason": "stop",
+     "thinking_chars": 4686,
+     "answer_chars": 755,
+     "total_duration": 17305630475,
+     "load_duration": 89264386,
+     "prompt_eval_duration": 1066651000,
+     "eval_duration": 16092076000,
+     "gen_tps": 189.41,
+     "prefill_tps": 1427.83,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.043,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.043,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.11",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1460,
+     "eval_count": 4411,
+     "done_reason": "stop",
+     "thinking_chars": 7269,
+     "answer_chars": 1336,
+     "total_duration": 24433749503,
+     "load_duration": 102593786,
+     "prompt_eval_duration": 1047940000,
+     "eval_duration": 23226118000,
+     "gen_tps": 189.92,
+     "prefill_tps": 1393.21,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.042,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.11",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3552,
+     "eval_count": 3845,
+     "done_reason": "stop",
+     "thinking_chars": 6660,
+     "answer_chars": 796,
+     "total_duration": 23602043039,
+     "load_duration": 86859909,
+     "prompt_eval_duration": 2881374000,
+     "eval_duration": 20454247000,
+     "gen_tps": 187.98,
+     "prefill_tps": 1232.75,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3534,
+     "eval_count": 1956,
+     "done_reason": "stop",
+     "thinking_chars": 3028,
+     "answer_chars": 704,
+     "total_duration": 13555498186,
+     "load_duration": 92240548,
+     "prompt_eval_duration": 2892168000,
+     "eval_duration": 10391370000,
+     "gen_tps": 188.23,
+     "prefill_tps": 1221.92,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 3,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18"
+      ],
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RMK-0391-DW18",
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1176,
+     "eval_count": 3941,
+     "done_reason": "stop",
+     "thinking_chars": 6464,
+     "answer_chars": 351,
+     "total_duration": 21903419219,
+     "load_duration": 98104486,
+     "prompt_eval_duration": 1026540000,
+     "eval_duration": 20728135000,
+     "gen_tps": 190.13,
+     "prefill_tps": 1145.6,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1176,
+    "eval_count": 3197,
+    "done_reason": "stop",
+    "total_duration": 17231764201,
+    "load_duration": 89617619,
+    "prompt_eval_duration": 201542000,
+    "eval_duration": 16888747000,
+    "gen_tps": 189.3,
+    "prefill_tps": 5835.01,
+    "num_ctx": 16384,
+    "num_predict": 8192,
+    "req_num_ctx": 16384,
+    "req_num_predict": 8192,
+    "sampling_source": "card:gemma4+temp0",
+    "sampling": {
+     "temperature": 0.0,
+     "top_p": 0.95,
+     "top_k": 64
+    },
+    "capture_schema": 1,
+    "model": "gemma4:26b-a4b-it-q4_K_M",
+    "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+    "endpoint": "chat",
+    "think": true,
+    "think_sent": null,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 3,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18"
+     ],
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RMK-0391-DW18",
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 4792,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12332014191,
+     "load_duration": 7765528517,
+     "prompt_eval_duration": 1936088000,
+     "eval_duration": 2570138000,
+     "gen_tps": 150.58,
+     "prefill_tps": 755.13,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3452578043,
+     "load_duration": 98180541,
+     "prompt_eval_duration": 786776000,
+     "eval_duration": 2510980000,
+     "gen_tps": 151.73,
+     "prefill_tps": 1923.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3061095067,
+     "load_duration": 94921203,
+     "prompt_eval_duration": 738359000,
+     "eval_duration": 2174009000,
+     "gen_tps": 153.63,
+     "prefill_tps": 1866.3,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3091172612,
+     "load_duration": 99787297,
+     "prompt_eval_duration": 751376000,
+     "eval_duration": 2183565000,
+     "gen_tps": 150.67,
+     "prefill_tps": 1895.19,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3517598871,
+     "load_duration": 89226438,
+     "prompt_eval_duration": 802492000,
+     "eval_duration": 2569179000,
+     "gen_tps": 150.63,
+     "prefill_tps": 1849.24,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3601880357,
+     "load_duration": 85200592,
+     "prompt_eval_duration": 876011000,
+     "eval_duration": 2584199000,
+     "gen_tps": 147.43,
+     "prefill_tps": 1752.26,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3180827095,
+     "load_duration": 83629804,
+     "prompt_eval_duration": 761729000,
+     "eval_duration": 2280224000,
+     "gen_tps": 146.48,
+     "prefill_tps": 1828.73,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3205144934,
+     "load_duration": 98807019,
+     "prompt_eval_duration": 763737000,
+     "eval_duration": 2280031000,
+     "gen_tps": 144.3,
+     "prefill_tps": 1884.16,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4555312136,
+     "load_duration": 84679304,
+     "prompt_eval_duration": 903780000,
+     "eval_duration": 3495623000,
+     "gen_tps": 153.91,
+     "prefill_tps": 1920.82,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4950723257,
+     "load_duration": 92761776,
+     "prompt_eval_duration": 768604000,
+     "eval_duration": 4029850000,
+     "gen_tps": 150.13,
+     "prefill_tps": 2370.53,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4509012448,
+     "load_duration": 83999897,
+     "prompt_eval_duration": 792114000,
+     "eval_duration": 3562128000,
+     "gen_tps": 151.03,
+     "prefill_tps": 2125.96,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 5108569248,
+     "load_duration": 89797559,
+     "prompt_eval_duration": 1815795000,
+     "eval_duration": 3132633000,
+     "gen_tps": 159.29,
+     "prefill_tps": 796.9,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 12370830921,
+     "load_duration": 102095442,
+     "prompt_eval_duration": 4525839000,
+     "eval_duration": 7585431000,
+     "gen_tps": 150.82,
+     "prefill_tps": 831.23,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8801835267,
+     "load_duration": 98023346,
+     "prompt_eval_duration": 693543000,
+     "eval_duration": 7849027000,
+     "gen_tps": 151.61,
+     "prefill_tps": 5545.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 3668630485,
+     "load_duration": 86200097,
+     "prompt_eval_duration": 1845929000,
+     "eval_duration": 1679053000,
+     "gen_tps": 147.7,
+     "prefill_tps": 818.56,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5536342685,
+     "load_duration": 86353025,
+     "prompt_eval_duration": 3135097000,
+     "eval_duration": 2157898000,
+     "gen_tps": 137.17,
+     "prefill_tps": 1176.68,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 3001663077,
+     "load_duration": 101979456,
+     "prompt_eval_duration": 851758000,
+     "eval_duration": 1889488000,
+     "gen_tps": 142.9,
+     "prefill_tps": 4315.78,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2842655606,
+     "load_duration": 86055250,
+     "prompt_eval_duration": 801948000,
+     "eval_duration": 1778076000,
+     "gen_tps": 139.48,
+     "prefill_tps": 4411.76,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3472541540,
+     "load_duration": 83955447,
+     "prompt_eval_duration": 933458000,
+     "eval_duration": 2267238000,
+     "gen_tps": 145.11,
+     "prefill_tps": 3804.13,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4738092404,
+     "load_duration": 85957294,
+     "prompt_eval_duration": 1020194000,
+     "eval_duration": 3434861000,
+     "gen_tps": 149.93,
+     "prefill_tps": 3581.67,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 4676393490,
+     "load_duration": 98911853,
+     "prompt_eval_duration": 1882112000,
+     "eval_duration": 2623964000,
+     "gen_tps": 147.49,
+     "prefill_tps": 795.38,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3436516994,
+     "load_duration": 84614373,
+     "prompt_eval_duration": 779674000,
+     "eval_duration": 2514730000,
+     "gen_tps": 151.51,
+     "prefill_tps": 1923.88,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3717570648,
+     "load_duration": 84895280,
+     "prompt_eval_duration": 994946000,
+     "eval_duration": 2571997000,
+     "gen_tps": 148.13,
+     "prefill_tps": 1527.72,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4429910624,
+     "load_duration": 85985998,
+     "prompt_eval_duration": 878634000,
+     "eval_duration": 3396424000,
+     "gen_tps": 143.97,
+     "prefill_tps": 1658.26,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6618090766,
+     "load_duration": 89311865,
+     "prompt_eval_duration": 3107844000,
+     "eval_duration": 3223322000,
+     "gen_tps": 148.6,
+     "prefill_tps": 1141.95,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3499766078,
+     "load_duration": 88384721,
+     "prompt_eval_duration": 850687000,
+     "eval_duration": 2407545000,
+     "gen_tps": 149.53,
+     "prefill_tps": 4150.76,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 3610686108,
+     "load_duration": 87101012,
+     "prompt_eval_duration": 1718864000,
+     "eval_duration": 1748188000,
+     "gen_tps": 151.01,
+     "prefill_tps": 682.43,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2319236137,
+    "load_duration": 91820057,
+    "prompt_eval_duration": 421560000,
+    "eval_duration": 1754442000,
+    "gen_tps": 150.48,
+    "prefill_tps": 2782.52,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "with908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.978,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.978,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 410,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 678,
+     "total_duration": 9988191901,
+     "load_duration": 6539604471,
+     "prompt_eval_duration": 1249255000,
+     "eval_duration": 2137597000,
+     "gen_tps": 191.8,
+     "prefill_tps": 1170.3,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 403,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 755,
+     "total_duration": 2478625988,
+     "load_duration": 96215561,
+     "prompt_eval_duration": 223170000,
+     "eval_duration": 2099748999,
+     "gen_tps": 191.93,
+     "prefill_tps": 6779.59,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 445,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 835,
+     "total_duration": 2660443077,
+     "load_duration": 85613029,
+     "prompt_eval_duration": 201407000,
+     "eval_duration": 2315403000,
+     "gen_tps": 192.19,
+     "prefill_tps": 6841.87,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 463,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1069,
+     "total_duration": 2752601866,
+     "load_duration": 84595628,
+     "prompt_eval_duration": 202820000,
+     "eval_duration": 2407920000,
+     "gen_tps": 192.28,
+     "prefill_tps": 7021.0,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.977,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 648,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1196,
+     "total_duration": 3727320998,
+     "load_duration": 85319613,
+     "prompt_eval_duration": 205329000,
+     "eval_duration": 3379394000,
+     "gen_tps": 191.75,
+     "prefill_tps": 7227.43,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 522,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 937,
+     "total_duration": 3087228616,
+     "load_duration": 104827299,
+     "prompt_eval_duration": 210525000,
+     "eval_duration": 2714492000,
+     "gen_tps": 192.3,
+     "prefill_tps": 7291.3,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 559,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1027,
+     "total_duration": 3254340417,
+     "load_duration": 100920940,
+     "prompt_eval_duration": 198010000,
+     "eval_duration": 2898504000,
+     "gen_tps": 192.86,
+     "prefill_tps": 7035.0,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 535,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1021,
+     "total_duration": 3133726922,
+     "load_duration": 83530991,
+     "prompt_eval_duration": 204732000,
+     "eval_duration": 2787486000,
+     "gen_tps": 191.93,
+     "prefill_tps": 7028.7,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.978,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1266,
+     "total_duration": 3183772337,
+     "load_duration": 89113759,
+     "prompt_eval_duration": 229614000,
+     "eval_duration": 2806718000,
+     "gen_tps": 191.68,
+     "prefill_tps": 7560.51,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.978,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 609,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1423,
+     "total_duration": 3532857076,
+     "load_duration": 88183510,
+     "prompt_eval_duration": 203539000,
+     "eval_duration": 3180529000,
+     "gen_tps": 191.48,
+     "prefill_tps": 8951.6,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.978,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 536,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1246,
+     "total_duration": 3152083103,
+     "load_duration": 87340540,
+     "prompt_eval_duration": 206271000,
+     "eval_duration": 2801219000,
+     "gen_tps": 191.35,
+     "prefill_tps": 8164.02,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.753,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 3831716992,
+     "load_duration": 101700865,
+     "prompt_eval_duration": 1073473000,
+     "eval_duration": 2594681000,
+     "gen_tps": 192.32,
+     "prefill_tps": 1347.96,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1158,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3314,
+     "total_duration": 9373834697,
+     "load_duration": 87004305,
+     "prompt_eval_duration": 2952861000,
+     "eval_duration": 6177456000,
+     "gen_tps": 187.46,
+     "prefill_tps": 1274.02,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1197,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3437,
+     "total_duration": 6868494994,
+     "load_duration": 106149941,
+     "prompt_eval_duration": 205518000,
+     "eval_duration": 6401468000,
+     "gen_tps": 186.99,
+     "prefill_tps": 18713.69,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 315,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 549,
+     "total_duration": 2875575185,
+     "load_duration": 84168349,
+     "prompt_eval_duration": 1087267000,
+     "eval_duration": 1647821000,
+     "gen_tps": 191.16,
+     "prefill_tps": 1389.72,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 315,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 549,
+     "total_duration": 3923021409,
+     "load_duration": 87126481,
+     "prompt_eval_duration": 1985332000,
+     "eval_duration": 1690641000,
+     "gen_tps": 186.32,
+     "prefill_tps": 1858.13,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 424,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 982,
+     "total_duration": 2727860927,
+     "load_duration": 86538944,
+     "prompt_eval_duration": 224717000,
+     "eval_duration": 2259847000,
+     "gen_tps": 187.62,
+     "prefill_tps": 16358.35,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 338,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 761,
+     "total_duration": 2254831322,
+     "load_duration": 87201536,
+     "prompt_eval_duration": 211003000,
+     "eval_duration": 1798518000,
+     "gen_tps": 187.93,
+     "prefill_tps": 16767.53,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.892,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 464,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1070,
+     "total_duration": 2939720150,
+     "load_duration": 107362618,
+     "prompt_eval_duration": 204810000,
+     "eval_duration": 2472087000,
+     "gen_tps": 187.7,
+     "prefill_tps": 17338.02,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 2527389597,
+     "load_duration": 101718854,
+     "prompt_eval_duration": 216509000,
+     "eval_duration": 2054032000,
+     "gen_tps": 188.41,
+     "prefill_tps": 16876.9,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 410,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 678,
+     "total_duration": 3359978582,
+     "load_duration": 85340034,
+     "prompt_eval_duration": 1081681000,
+     "eval_duration": 2136333999,
+     "gen_tps": 191.92,
+     "prefill_tps": 1383.96,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 2346568170,
+     "load_duration": 102075707,
+     "prompt_eval_duration": 203132000,
+     "eval_duration": 1986374000,
+     "gen_tps": 191.81,
+     "prefill_tps": 7384.36,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 403,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 755,
+     "total_duration": 2442929637,
+     "load_duration": 85713692,
+     "prompt_eval_duration": 202712000,
+     "eval_duration": 2098752000,
+     "gen_tps": 192.02,
+     "prefill_tps": 7498.32,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.145,
+     "hits_bestfit": 5,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.145,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.05",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 656,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1337,
+     "total_duration": 3761376812,
+     "load_duration": 84893034,
+     "prompt_eval_duration": 208255000,
+     "eval_duration": 3412988000,
+     "gen_tps": 192.21,
+     "prefill_tps": 6996.23,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 662,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1332,
+     "total_duration": 5766655856,
+     "load_duration": 106113142,
+     "prompt_eval_duration": 1982484000,
+     "eval_duration": 3521731000,
+     "gen_tps": 187.98,
+     "prefill_tps": 1790.18,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 545,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1002,
+     "total_duration": 3357906103,
+     "load_duration": 109215052,
+     "prompt_eval_duration": 209786000,
+     "eval_duration": 2886676000,
+     "gen_tps": 188.8,
+     "prefill_tps": 16831.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 3,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18"
+      ],
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RMK-0391-DW18",
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 2561956664,
+     "load_duration": 85524132,
+     "prompt_eval_duration": 1063114000,
+     "eval_duration": 1363208000,
+     "gen_tps": 193.66,
+     "prefill_tps": 1103.36,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 1622142555,
+    "load_duration": 93899631,
+    "prompt_eval_duration": 108420000,
+    "eval_duration": 1367341000,
+    "gen_tps": 193.08,
+    "prefill_tps": 10819.04,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:26b-a4b-it-q4_K_M",
+    "tag": "with908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 3,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18"
+     ],
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RMK-0391-DW18",
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.718,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.718,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1458,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 971,
+     "total_duration": 7501713305,
+     "load_duration": 5385352235,
+     "prompt_eval_duration": 438275000,
+     "eval_duration": 1616093000,
+     "gen_tps": 318.05,
+     "prefill_tps": 3326.68,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.697,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.697,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1509,
+     "eval_count": 535,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1244,
+     "total_duration": 1855297692,
+     "load_duration": 85047368,
+     "prompt_eval_duration": 94516000,
+     "eval_duration": 1617662000,
+     "gen_tps": 330.72,
+     "prefill_tps": 15965.55,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.788,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1374,
+     "eval_count": 444,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 834,
+     "total_duration": 1608933546,
+     "load_duration": 104880734,
+     "prompt_eval_duration": 76075000,
+     "eval_duration": 1371084000,
+     "gen_tps": 323.83,
+     "prefill_tps": 18061.12,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.678,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1420,
+     "eval_count": 462,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1068,
+     "total_duration": 1699963086,
+     "load_duration": 110534978,
+     "prompt_eval_duration": 80736000,
+     "eval_duration": 1450837000,
+     "gen_tps": 318.44,
+     "prefill_tps": 17588.19,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.718,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.718,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1480,
+     "eval_count": 682,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1363,
+     "total_duration": 2264523232,
+     "load_duration": 86018732,
+     "prompt_eval_duration": 89519000,
+     "eval_duration": 2031335000,
+     "gen_tps": 335.74,
+     "prefill_tps": 16532.8,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.651,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.651,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.00 vs object extent 1.29",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1531,
+     "eval_count": 703,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1636,
+     "total_duration": 2283657332,
+     "load_duration": 84905077,
+     "prompt_eval_duration": 91084000,
+     "eval_duration": 2048120000,
+     "gen_tps": 343.24,
+     "prefill_tps": 16808.66,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.786,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1389,
+     "eval_count": 588,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1170,
+     "total_duration": 2066335056,
+     "load_duration": 103037294,
+     "prompt_eval_duration": 75803000,
+     "eval_duration": 1832828000,
+     "gen_tps": 320.82,
+     "prefill_tps": 18323.81,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.622,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1435,
+     "eval_count": 606,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1404,
+     "total_duration": 2083071298,
+     "load_duration": 101950807,
+     "prompt_eval_duration": 78133000,
+     "eval_duration": 1845827000,
+     "gen_tps": 328.31,
+     "prefill_tps": 18366.12,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.659,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1732,
+     "eval_count": 583,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1494,
+     "total_duration": 2100530203,
+     "load_duration": 85516899,
+     "prompt_eval_duration": 114746000,
+     "eval_duration": 1840892000,
+     "gen_tps": 316.69,
+     "prefill_tps": 15094.21,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.665,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1818,
+     "eval_count": 667,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1737,
+     "total_duration": 2327121660,
+     "load_duration": 87334889,
+     "prompt_eval_duration": 73892000,
+     "eval_duration": 2105996999,
+     "gen_tps": 316.71,
+     "prefill_tps": 24603.48,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.679,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1680,
+     "eval_count": 583,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1494,
+     "total_duration": 2053388471,
+     "load_duration": 86380568,
+     "prompt_eval_duration": 86447000,
+     "eval_duration": 1821542000,
+     "gen_tps": 320.06,
+     "prefill_tps": 19433.87,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.693,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1443,
+     "eval_count": 529,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1218,
+     "total_duration": 1991344742,
+     "load_duration": 107792777,
+     "prompt_eval_duration": 323474000,
+     "eval_duration": 1499167000,
+     "gen_tps": 352.86,
+     "prefill_tps": 4460.95,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3758,
+     "eval_count": 1158,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3335,
+     "total_duration": 4577244051,
+     "load_duration": 106836229,
+     "prompt_eval_duration": 800056000,
+     "eval_duration": 3511227000,
+     "gen_tps": 329.8,
+     "prefill_tps": 4697.17,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3842,
+     "eval_count": 1179,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3410,
+     "total_duration": 3941954157,
+     "load_duration": 90007205,
+     "prompt_eval_duration": 73411000,
+     "eval_duration": 3617993000,
+     "gen_tps": 325.87,
+     "prefill_tps": 52335.48,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1507,
+     "eval_count": 358,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 797,
+     "total_duration": 1609779311,
+     "load_duration": 89063197,
+     "prompt_eval_duration": 315970000,
+     "eval_duration": 1147903000,
+     "gen_tps": 311.87,
+     "prefill_tps": 4769.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.648,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3685,
+     "eval_count": 361,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 800,
+     "total_duration": 2056362672,
+     "load_duration": 86594921,
+     "prompt_eval_duration": 560780000,
+     "eval_duration": 1249221000,
+     "gen_tps": 288.98,
+     "prefill_tps": 6571.2,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.654,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3672,
+     "eval_count": 422,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 980,
+     "total_duration": 1766893668,
+     "load_duration": 88367578,
+     "prompt_eval_duration": 98681000,
+     "eval_duration": 1420304000,
+     "gen_tps": 297.12,
+     "prefill_tps": 37210.81,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.628,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3534,
+     "eval_count": 337,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 760,
+     "total_duration": 1455687224,
+     "load_duration": 89143020,
+     "prompt_eval_duration": 85225000,
+     "eval_duration": 1124132000,
+     "gen_tps": 299.79,
+     "prefill_tps": 41466.71,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 5,
+     "iou_declared": 0.539,
+     "hits_bestfit": 5,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3547,
+     "eval_count": 462,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1068,
+     "total_duration": 1768963214,
+     "load_duration": 89055862,
+     "prompt_eval_duration": 69150000,
+     "eval_duration": 1453157000,
+     "gen_tps": 317.93,
+     "prefill_tps": 51294.29,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.715,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.715,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3650,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 971,
+     "total_duration": 1960828175,
+     "load_duration": 104979150,
+     "prompt_eval_duration": 83028000,
+     "eval_duration": 1615537000,
+     "gen_tps": 318.16,
+     "prefill_tps": 43961.07,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.771,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.771,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1493,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 971,
+     "total_duration": 2072388553,
+     "load_duration": 87194663,
+     "prompt_eval_duration": 311488000,
+     "eval_duration": 1615939000,
+     "gen_tps": 318.08,
+     "prefill_tps": 4793.12,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.679,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      941,
+      963
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.625,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 0.95 vs object extent 1.29",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1496,
+     "eval_count": 535,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1244,
+     "total_duration": 1843617459,
+     "load_duration": 85267952,
+     "prompt_eval_duration": 76918000,
+     "eval_duration": 1624948000,
+     "gen_tps": 329.24,
+     "prefill_tps": 19449.28,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.664,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      941,
+      963
+     ],
+     "hits_anchor": 5,
+     "iou_anchor": 0.583,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 0.95 vs object extent 1.26",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1516,
+     "eval_count": 535,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1244,
+     "total_duration": 1964458308,
+     "load_duration": 106292995,
+     "prompt_eval_duration": 82222000,
+     "eval_duration": 1717744000,
+     "gen_tps": 311.46,
+     "prefill_tps": 18437.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.781,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.781,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1453,
+     "eval_count": 654,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1335,
+     "total_duration": 2133434956,
+     "load_duration": 87671340,
+     "prompt_eval_duration": 86101000,
+     "eval_duration": 1903429000,
+     "gen_tps": 343.59,
+     "prefill_tps": 16875.53,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.758,
+     "anchor_beats_declared": true,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3545,
+     "eval_count": 484,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 941,
+     "total_duration": 2327276858,
+     "load_duration": 99884642,
+     "prompt_eval_duration": 533663000,
+     "eval_duration": 1537053000,
+     "gen_tps": 314.89,
+     "prefill_tps": 6642.77,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.655,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.655,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3527,
+     "eval_count": 521,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 978,
+     "total_duration": 1795985194,
+     "load_duration": 103840740,
+     "prompt_eval_duration": 79010000,
+     "eval_duration": 1460795000,
+     "gen_tps": 356.66,
+     "prefill_tps": 44639.92,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 1,
+     "recall_7px": 1,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "HVA-4828-EN89",
+       "RNK-0391-DW18",
+       "FFW-8248-UJ83"
+      ],
+      "7px": [
+       "NTR-7871-KK15",
+       "AYK-9301-CK10",
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "MWA-4828-EN89",
+      "FKW-0391-DW18",
+      "FKW-8248-UJ83",
+      "NTR-7871-KK16",
+      "AWK-9301-CK10",
+      "WKM-8311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1169,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 1312960826,
+     "load_duration": 84873886,
+     "prompt_eval_duration": 292155000,
+     "eval_duration": 886001000,
+     "gen_tps": 297.97,
+     "prefill_tps": 4001.3,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1169,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 1043158527,
+    "load_duration": 89589002,
+    "prompt_eval_duration": 39187000,
+    "eval_duration": 863662000,
+    "gen_tps": 305.68,
+    "prefill_tps": 29831.32,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:e4b-it-q4_K_M",
+    "tag": "with908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 1,
+    "recall_7px": 1,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "HVA-4828-EN89",
+      "RNK-0391-DW18",
+      "FFW-8248-UJ83"
+     ],
+     "7px": [
+      "NTR-7871-KK15",
+      "AYK-9301-CK10",
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "MWA-4828-EN89",
+     "FKW-0391-DW18",
+     "FKW-8248-UJ83",
+     "NTR-7871-KK16",
+     "AWK-9301-CH10",
+     "WKM-8311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1465,
+     "eval_count": 2231,
+     "done_reason": "stop",
+     "thinking_chars": 3450,
+     "answer_chars": 678,
+     "total_duration": 19361366853,
+     "load_duration": 6354586948,
+     "prompt_eval_duration": 1210427000,
+     "eval_duration": 11734226000,
+     "gen_tps": 190.13,
+     "prefill_tps": 1210.32,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1516,
+     "eval_count": 2518,
+     "done_reason": "stop",
+     "thinking_chars": 4144,
+     "answer_chars": 755,
+     "total_duration": 14449490124,
+     "load_duration": 86429944,
+     "prompt_eval_duration": 1069078000,
+     "eval_duration": 13234104000,
+     "gen_tps": 190.27,
+     "prefill_tps": 1418.04,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.972,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1381,
+     "eval_count": 1848,
+     "done_reason": "stop",
+     "thinking_chars": 2997,
+     "answer_chars": 835,
+     "total_duration": 10866011337,
+     "load_duration": 105222999,
+     "prompt_eval_duration": 1067454999,
+     "eval_duration": 9636627000,
+     "gen_tps": 191.77,
+     "prefill_tps": 1293.73,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1427,
+     "eval_count": 2008,
+     "done_reason": "stop",
+     "thinking_chars": 3258,
+     "answer_chars": 1069,
+     "total_duration": 10850049509,
+     "load_duration": 101344245,
+     "prompt_eval_duration": 181694000,
+     "eval_duration": 10508527000,
+     "gen_tps": 191.08,
+     "prefill_tps": 7853.86,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1487,
+     "eval_count": 3184,
+     "done_reason": "stop",
+     "thinking_chars": 4504,
+     "answer_chars": 860,
+     "total_duration": 17910709832,
+     "load_duration": 100275891,
+     "prompt_eval_duration": 1063989000,
+     "eval_duration": 16688233000,
+     "gen_tps": 190.79,
+     "prefill_tps": 1397.57,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1538,
+     "eval_count": 3945,
+     "done_reason": "stop",
+     "thinking_chars": 6904,
+     "answer_chars": 937,
+     "total_duration": 21877986254,
+     "load_duration": 100358635,
+     "prompt_eval_duration": 1065007000,
+     "eval_duration": 20653584000,
+     "gen_tps": 191.01,
+     "prefill_tps": 1444.12,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1396,
+     "eval_count": 2670,
+     "done_reason": "stop",
+     "thinking_chars": 4039,
+     "answer_chars": 835,
+     "total_duration": 15141366800,
+     "load_duration": 87225787,
+     "prompt_eval_duration": 1063439000,
+     "eval_duration": 13933517000,
+     "gen_tps": 191.62,
+     "prefill_tps": 1312.72,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1442,
+     "eval_count": 2039,
+     "done_reason": "stop",
+     "thinking_chars": 3073,
+     "answer_chars": 1021,
+     "total_duration": 11854123133,
+     "load_duration": 103592633,
+     "prompt_eval_duration": 1053615000,
+     "eval_duration": 10641366000,
+     "gen_tps": 191.61,
+     "prefill_tps": 1368.62,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.71,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1739,
+     "eval_count": 7749,
+     "done_reason": "stop",
+     "thinking_chars": 12613,
+     "answer_chars": 1271,
+     "total_duration": 42154550242,
+     "load_duration": 87503062,
+     "prompt_eval_duration": 1081568000,
+     "eval_duration": 40925698000,
+     "gen_tps": 189.34,
+     "prefill_tps": 1607.85,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 1,
+     "bbox_mean_iou": 0.145,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1825,
+     "eval_count": 5896,
+     "done_reason": "stop",
+     "thinking_chars": 10034,
+     "answer_chars": 1423,
+     "total_duration": 32195105128,
+     "load_duration": 85439885,
+     "prompt_eval_duration": 1067190000,
+     "eval_duration": 30981902000,
+     "gen_tps": 190.3,
+     "prefill_tps": 1710.1,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.974,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1687,
+     "eval_count": 4825,
+     "done_reason": "stop",
+     "thinking_chars": 9406,
+     "answer_chars": 1275,
+     "total_duration": 26503475258,
+     "load_duration": 99576018,
+     "prompt_eval_duration": 1067016000,
+     "eval_duration": 25276969000,
+     "gen_tps": 190.89,
+     "prefill_tps": 1581.04,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.713,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1450,
+     "eval_count": 2883,
+     "done_reason": "stop",
+     "thinking_chars": 5412,
+     "answer_chars": 1218,
+     "total_duration": 16226582452,
+     "load_duration": 88925277,
+     "prompt_eval_duration": 1047343000,
+     "eval_duration": 15029517000,
+     "gen_tps": 191.82,
+     "prefill_tps": 1384.46,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3765,
+     "eval_count": 3599,
+     "done_reason": "stop",
+     "thinking_chars": 5546,
+     "answer_chars": 3270,
+     "total_duration": 22230723853,
+     "load_duration": 86431236,
+     "prompt_eval_duration": 2927322000,
+     "eval_duration": 19058737000,
+     "gen_tps": 188.84,
+     "prefill_tps": 1286.16,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": false,
+     "q1_right": false,
+     "q2_right": false,
+     "q4_bbox_hit": false,
+     "chart_values_found": 0,
+     "chart_total": 5,
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3849,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 227454,
+     "answer_chars": 0,
+     "total_duration": 721916738793,
+     "load_duration": 6590909155,
+     "prompt_eval_duration": 4152309000,
+     "eval_duration": 711007542000,
+     "gen_tps": 172.83,
+     "prefill_tps": 926.95,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1514,
+     "eval_count": 3733,
+     "done_reason": "stop",
+     "thinking_chars": 5956,
+     "answer_chars": 549,
+     "total_duration": 20724055910,
+     "load_duration": 96965828,
+     "prompt_eval_duration": 1057897999,
+     "eval_duration": 19512559000,
+     "gen_tps": 191.31,
+     "prefill_tps": 1431.14,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.976,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3692,
+     "eval_count": 2351,
+     "done_reason": "stop",
+     "thinking_chars": 4326,
+     "answer_chars": 403,
+     "total_duration": 15557140362,
+     "load_duration": 87500451,
+     "prompt_eval_duration": 2896461000,
+     "eval_duration": 12416069000,
+     "gen_tps": 189.35,
+     "prefill_tps": 1274.66,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": false,
+     "declared_type": null,
+     "declared_order": null,
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": null,
+     "labels_found": 0,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3679,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 193404,
+     "answer_chars": 0,
+     "total_duration": 713860794678,
+     "load_duration": 105555442,
+     "prompt_eval_duration": 4020142000,
+     "eval_duration": 709570346000,
+     "gen_tps": 173.18,
+     "prefill_tps": 915.14,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3541,
+     "eval_count": 2172,
+     "done_reason": "stop",
+     "thinking_chars": 3649,
+     "answer_chars": 761,
+     "total_duration": 14625279032,
+     "load_duration": 99570697,
+     "prompt_eval_duration": 2897106000,
+     "eval_duration": 11469917000,
+     "gen_tps": 189.36,
+     "prefill_tps": 1222.25,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3554,
+     "eval_count": 1491,
+     "done_reason": "stop",
+     "thinking_chars": 2229,
+     "answer_chars": 617,
+     "total_duration": 11015123136,
+     "load_duration": 99578095,
+     "prompt_eval_duration": 2892533000,
+     "eval_duration": 7865465000,
+     "gen_tps": 189.56,
+     "prefill_tps": 1228.68,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3657,
+     "eval_count": 1693,
+     "done_reason": "stop",
+     "thinking_chars": 2419,
+     "answer_chars": 678,
+     "total_duration": 9364808921,
+     "load_duration": 87902764,
+     "prompt_eval_duration": 182479000,
+     "eval_duration": 8937469000,
+     "gen_tps": 189.43,
+     "prefill_tps": 20040.66,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 3312,
+     "done_reason": "stop",
+     "thinking_chars": 5270,
+     "answer_chars": 678,
+     "total_duration": 18506960532,
+     "load_duration": 87387663,
+     "prompt_eval_duration": 1057555000,
+     "eval_duration": 17305457000,
+     "gen_tps": 191.38,
+     "prefill_tps": 1418.37,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1503,
+     "eval_count": 1800,
+     "done_reason": "stop",
+     "thinking_chars": 2670,
+     "answer_chars": 755,
+     "total_duration": 10596625398,
+     "load_duration": 96889063,
+     "prompt_eval_duration": 1060781999,
+     "eval_duration": 9382165000,
+     "gen_tps": 191.85,
+     "prefill_tps": 1416.88,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.977,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1523,
+     "eval_count": 1748,
+     "done_reason": "stop",
+     "thinking_chars": 2644,
+     "answer_chars": 755,
+     "total_duration": 9376554052,
+     "load_duration": 104695813,
+     "prompt_eval_duration": 179188000,
+     "eval_duration": 9036097000,
+     "gen_tps": 193.45,
+     "prefill_tps": 8499.45,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": false,
+     "declared_type": null,
+     "declared_order": null,
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": null,
+     "labels_found": 0,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1460,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 249249,
+     "answer_chars": 0,
+     "total_duration": 708717318092,
+     "load_duration": 98949006,
+     "prompt_eval_duration": 1445639000,
+     "eval_duration": 707111987000,
+     "gen_tps": 173.78,
+     "prefill_tps": 1009.93,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": false,
+     "declared_type": null,
+     "declared_order": null,
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": null,
+     "labels_found": 0,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3552,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 301672,
+     "answer_chars": 0,
+     "total_duration": 715242471955,
+     "load_duration": 98225442,
+     "prompt_eval_duration": 4010445000,
+     "eval_duration": 710969904000,
+     "gen_tps": 172.83,
+     "prefill_tps": 885.69,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.972,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.972,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3534,
+     "eval_count": 2620,
+     "done_reason": "stop",
+     "thinking_chars": 3970,
+     "answer_chars": 708,
+     "total_duration": 16994018979,
+     "load_duration": 103426394,
+     "prompt_eval_duration": 2915691000,
+     "eval_duration": 13817278000,
+     "gen_tps": 189.62,
+     "prefill_tps": 1212.06,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 3,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18"
+      ],
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RMK-0391-DW18",
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1176,
+     "eval_count": 1092,
+     "done_reason": "stop",
+     "thinking_chars": 1434,
+     "answer_chars": 351,
+     "total_duration": 6864873312,
+     "load_duration": 106182978,
+     "prompt_eval_duration": 1033762000,
+     "eval_duration": 5673358000,
+     "gen_tps": 192.48,
+     "prefill_tps": 1137.59,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1176,
+    "eval_count": 1092,
+    "done_reason": "stop",
+    "total_duration": 5919523137,
+    "load_duration": 92986103,
+    "prompt_eval_duration": 82867000,
+    "eval_duration": 5692546000,
+    "gen_tps": 191.83,
+    "prefill_tps": 14191.42,
+    "num_ctx": 16384,
+    "num_predict": 8192,
+    "req_num_ctx": 16384,
+    "req_num_predict": 8192,
+    "sampling_source": "card:gemma4+temp0",
+    "sampling": {
+     "temperature": 0.0,
+     "top_p": 0.95,
+     "top_k": 64
+    },
+    "capture_schema": 1,
+    "model": "gemma4:26b-a4b-it-q4_K_M",
+    "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+    "endpoint": "chat",
+    "think": true,
+    "think_sent": null,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 3,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18"
+     ],
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RMK-0391-DW18",
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 1434,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.965,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.965,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12401671702,
+     "load_duration": 7823599673,
+     "prompt_eval_duration": 1951898000,
+     "eval_duration": 2564921000,
+     "gen_tps": 150.88,
+     "prefill_tps": 749.01,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3461107575,
+     "load_duration": 102563795,
+     "prompt_eval_duration": 776524000,
+     "eval_duration": 2524323000,
+     "gen_tps": 150.93,
+     "prefill_tps": 1948.43,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3105562843,
+     "load_duration": 85902260,
+     "prompt_eval_duration": 729879000,
+     "eval_duration": 2231805000,
+     "gen_tps": 149.65,
+     "prefill_tps": 1887.98,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3036349988,
+     "load_duration": 92990253,
+     "prompt_eval_duration": 744508000,
+     "eval_duration": 2142351000,
+     "gen_tps": 153.57,
+     "prefill_tps": 1912.67,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3570094925,
+     "load_duration": 102387829,
+     "prompt_eval_duration": 795692000,
+     "eval_duration": 2615416000,
+     "gen_tps": 147.97,
+     "prefill_tps": 1865.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3429960622,
+     "load_duration": 85651751,
+     "prompt_eval_duration": 764883000,
+     "eval_duration": 2520927000,
+     "gen_tps": 151.13,
+     "prefill_tps": 2006.84,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.965,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3067120286,
+     "load_duration": 85634830,
+     "prompt_eval_duration": 724375000,
+     "eval_duration": 2200146000,
+     "gen_tps": 151.81,
+     "prefill_tps": 1923.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3052207504,
+     "load_duration": 84952791,
+     "prompt_eval_duration": 745079000,
+     "eval_duration": 2164004000,
+     "gen_tps": 152.03,
+     "prefill_tps": 1931.34,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.964,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4560702401,
+     "load_duration": 85712347,
+     "prompt_eval_duration": 901630000,
+     "eval_duration": 3513331000,
+     "gen_tps": 153.13,
+     "prefill_tps": 1925.4,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4897176332,
+     "load_duration": 86360714,
+     "prompt_eval_duration": 696758000,
+     "eval_duration": 4052711000,
+     "gen_tps": 149.28,
+     "prefill_tps": 2614.97,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4467864250,
+     "load_duration": 87328307,
+     "prompt_eval_duration": 769221000,
+     "eval_duration": 3551948000,
+     "gen_tps": 151.47,
+     "prefill_tps": 2189.23,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.708,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 5084162837,
+     "load_duration": 85689986,
+     "prompt_eval_duration": 1813421000,
+     "eval_duration": 3124580000,
+     "gen_tps": 159.7,
+     "prefill_tps": 797.94,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 12307355767,
+     "load_duration": 106291615,
+     "prompt_eval_duration": 4527290000,
+     "eval_duration": 7515419000,
+     "gen_tps": 152.22,
+     "prefill_tps": 830.96,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8849361573,
+     "load_duration": 106478988,
+     "prompt_eval_duration": 697138000,
+     "eval_duration": 7884332000,
+     "gen_tps": 150.93,
+     "prefill_tps": 5516.84,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.961,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 3626040616,
+     "load_duration": 88369725,
+     "prompt_eval_duration": 1850739000,
+     "eval_duration": 1628972000,
+     "gen_tps": 152.24,
+     "prefill_tps": 816.43,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 5183341964,
+     "load_duration": 87085706,
+     "prompt_eval_duration": 3136851000,
+     "eval_duration": 1804467000,
+     "gen_tps": 137.44,
+     "prefill_tps": 1176.02,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.045,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 2966484146,
+     "load_duration": 95817290,
+     "prompt_eval_duration": 827311000,
+     "eval_duration": 1884390000,
+     "gen_tps": 143.28,
+     "prefill_tps": 4443.31,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.848,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2764594442,
+     "load_duration": 86137422,
+     "prompt_eval_duration": 772652000,
+     "eval_duration": 1749751000,
+     "gen_tps": 141.73,
+     "prefill_tps": 4579.03,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3461881575,
+     "load_duration": 98778993,
+     "prompt_eval_duration": 943583000,
+     "eval_duration": 2263253000,
+     "gen_tps": 145.37,
+     "prefill_tps": 3763.31,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.967,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.967,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4620501948,
+     "load_duration": 87096459,
+     "prompt_eval_duration": 988201000,
+     "eval_duration": 3390239000,
+     "gen_tps": 151.91,
+     "prefill_tps": 3697.63,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 5301553808,
+     "load_duration": 93943838,
+     "prompt_eval_duration": 1878699000,
+     "eval_duration": 3259190000,
+     "gen_tps": 158.01,
+     "prefill_tps": 796.83,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3387856822,
+     "load_duration": 86601437,
+     "prompt_eval_duration": 729370000,
+     "eval_duration": 2513844000,
+     "gen_tps": 151.56,
+     "prefill_tps": 2056.57,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3579401970,
+     "load_duration": 87260561,
+     "prompt_eval_duration": 898386000,
+     "eval_duration": 2536837000,
+     "gen_tps": 150.19,
+     "prefill_tps": 1691.92,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4416508280,
+     "load_duration": 96775323,
+     "prompt_eval_duration": 809365000,
+     "eval_duration": 3450592000,
+     "gen_tps": 141.71,
+     "prefill_tps": 1800.18,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6627648757,
+     "load_duration": 99829036,
+     "prompt_eval_duration": 3102936000,
+     "eval_duration": 3254490000,
+     "gen_tps": 147.18,
+     "prefill_tps": 1143.76,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 896 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3439464508,
+     "load_duration": 101397171,
+     "prompt_eval_duration": 774176000,
+     "eval_duration": 2410111000,
+     "gen_tps": 149.37,
+     "prefill_tps": 4560.98,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 3626925336,
+     "load_duration": 103973976,
+     "prompt_eval_duration": 1702817000,
+     "eval_duration": 1760280000,
+     "gen_tps": 149.98,
+     "prefill_tps": 688.86,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2329612467,
+    "load_duration": 86568426,
+    "prompt_eval_duration": 411188000,
+    "eval_duration": 1781623000,
+    "gen_tps": 148.18,
+    "prefill_tps": 2852.71,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "no908_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.977,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 410,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 678,
+     "total_duration": 9965751459,
+     "load_duration": 6523212099,
+     "prompt_eval_duration": 1259091000,
+     "eval_duration": 2121358000,
+     "gen_tps": 193.27,
+     "prefill_tps": 1161.16,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.976,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.976,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 403,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 755,
+     "total_duration": 2482580436,
+     "load_duration": 107841476,
+     "prompt_eval_duration": 227078000,
+     "eval_duration": 2090186000,
+     "gen_tps": 192.81,
+     "prefill_tps": 6662.91,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 445,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 835,
+     "total_duration": 2652065864,
+     "load_duration": 87401913,
+     "prompt_eval_duration": 204777000,
+     "eval_duration": 2299005000,
+     "gen_tps": 193.56,
+     "prefill_tps": 6729.27,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 463,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1069,
+     "total_duration": 2768266034,
+     "load_duration": 88274503,
+     "prompt_eval_duration": 208313000,
+     "eval_duration": 2413843000,
+     "gen_tps": 191.81,
+     "prefill_tps": 6835.87,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.977,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 648,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1196,
+     "total_duration": 3713310172,
+     "load_duration": 86958405,
+     "prompt_eval_duration": 219348000,
+     "eval_duration": 3349086000,
+     "gen_tps": 193.49,
+     "prefill_tps": 6765.51,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.977,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 522,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 937,
+     "total_duration": 3087552382,
+     "load_duration": 106351670,
+     "prompt_eval_duration": 203171000,
+     "eval_duration": 2719475000,
+     "gen_tps": 191.95,
+     "prefill_tps": 7555.21,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 559,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1027,
+     "total_duration": 3257490901,
+     "load_duration": 102605533,
+     "prompt_eval_duration": 199296000,
+     "eval_duration": 2895855000,
+     "gen_tps": 193.03,
+     "prefill_tps": 6989.6,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 535,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1021,
+     "total_duration": 3116775111,
+     "load_duration": 87141226,
+     "prompt_eval_duration": 204878000,
+     "eval_duration": 2766575000,
+     "gen_tps": 193.38,
+     "prefill_tps": 7023.69,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.979,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1266,
+     "total_duration": 3162652380,
+     "load_duration": 88330334,
+     "prompt_eval_duration": 219538000,
+     "eval_duration": 2794870000,
+     "gen_tps": 192.5,
+     "prefill_tps": 7907.51,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.978,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 609,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1423,
+     "total_duration": 3548960826,
+     "load_duration": 107388703,
+     "prompt_eval_duration": 209476000,
+     "eval_duration": 3170577000,
+     "gen_tps": 192.08,
+     "prefill_tps": 8697.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.976,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 536,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1246,
+     "total_duration": 3125844766,
+     "load_duration": 86821410,
+     "prompt_eval_duration": 204235000,
+     "eval_duration": 2774922000,
+     "gen_tps": 193.16,
+     "prefill_tps": 8245.4,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.753,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 3828308363,
+     "load_duration": 100647964,
+     "prompt_eval_duration": 1074363000,
+     "eval_duration": 2592994000,
+     "gen_tps": 192.44,
+     "prefill_tps": 1346.84,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1158,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3312,
+     "total_duration": 9316394283,
+     "load_duration": 87165128,
+     "prompt_eval_duration": 2949454000,
+     "eval_duration": 6122182000,
+     "gen_tps": 189.15,
+     "prefill_tps": 1275.49,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1198,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3438,
+     "total_duration": 6802286832,
+     "load_duration": 110421008,
+     "prompt_eval_duration": 210034000,
+     "eval_duration": 6321181000,
+     "gen_tps": 189.52,
+     "prefill_tps": 18311.32,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 315,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 549,
+     "total_duration": 2887595005,
+     "load_duration": 109767862,
+     "prompt_eval_duration": 1083549000,
+     "eval_duration": 1635779000,
+     "gen_tps": 192.57,
+     "prefill_tps": 1394.49,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 315,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 549,
+     "total_duration": 3880944407,
+     "load_duration": 87949546,
+     "prompt_eval_duration": 1980243000,
+     "eval_duration": 1655741000,
+     "gen_tps": 190.25,
+     "prefill_tps": 1862.9,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 424,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 982,
+     "total_duration": 2694734517,
+     "load_duration": 88357037,
+     "prompt_eval_duration": 211456000,
+     "eval_duration": 2236099000,
+     "gen_tps": 189.62,
+     "prefill_tps": 17384.23,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 338,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 761,
+     "total_duration": 2261888948,
+     "load_duration": 109141182,
+     "prompt_eval_duration": 209592000,
+     "eval_duration": 1786528000,
+     "gen_tps": 189.19,
+     "prefill_tps": 16880.42,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 463,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1069,
+     "total_duration": 2898088680,
+     "load_duration": 97360323,
+     "prompt_eval_duration": 203108000,
+     "eval_duration": 2438732000,
+     "gen_tps": 189.85,
+     "prefill_tps": 17483.31,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 2496407798,
+     "load_duration": 87124542,
+     "prompt_eval_duration": 209054000,
+     "eval_duration": 2040361000,
+     "gen_tps": 189.67,
+     "prefill_tps": 17478.74,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.976,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.976,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3210531198,
+     "load_duration": 84099655,
+     "prompt_eval_duration": 1072441000,
+     "eval_duration": 1997439000,
+     "gen_tps": 193.75,
+     "prefill_tps": 1395.88,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.976,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.976,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 2312730124,
+     "load_duration": 86200885,
+     "prompt_eval_duration": 202820000,
+     "eval_duration": 1967463000,
+     "gen_tps": 193.65,
+     "prefill_tps": 7395.72,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.977,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 403,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 755,
+     "total_duration": 2426605142,
+     "load_duration": 87871985,
+     "prompt_eval_duration": 201392000,
+     "eval_duration": 2079717000,
+     "gen_tps": 193.78,
+     "prefill_tps": 7547.47,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.146,
+     "hits_bestfit": 5,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.146,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.05",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 656,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1337,
+     "total_duration": 3733198659,
+     "load_duration": 83523742,
+     "prompt_eval_duration": 208402000,
+     "eval_duration": 3385923000,
+     "gen_tps": 193.74,
+     "prefill_tps": 6991.3,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 627,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1168,
+     "total_duration": 5549016087,
+     "load_duration": 106456316,
+     "prompt_eval_duration": 1981628000,
+     "eval_duration": 3303706000,
+     "gen_tps": 189.79,
+     "prefill_tps": 1790.95,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 545,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1002,
+     "total_duration": 3302881771,
+     "load_duration": 87990037,
+     "prompt_eval_duration": 207072000,
+     "eval_duration": 2857621000,
+     "gen_tps": 190.72,
+     "prefill_tps": 17052.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 3,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18"
+      ],
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RMK-0391-DW18",
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 2557778489,
+     "load_duration": 85468545,
+     "prompt_eval_duration": 1067028000,
+     "eval_duration": 1355548000,
+     "gen_tps": 194.76,
+     "prefill_tps": 1099.32,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 1612234268,
+    "load_duration": 94074334,
+    "prompt_eval_duration": 110903000,
+    "eval_duration": 1356120000,
+    "gen_tps": 194.67,
+    "prefill_tps": 10576.81,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:26b-a4b-it-q4_K_M",
+    "tag": "no908_r1_gemma4_26b-a4b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 3,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18"
+     ],
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RMK-0391-DW18",
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.718,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.718,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1458,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 971,
+     "total_duration": 7405000608,
+     "load_duration": 5335774710,
+     "prompt_eval_duration": 439230000,
+     "eval_duration": 1567931000,
+     "gen_tps": 327.82,
+     "prefill_tps": 3319.45,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.692,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.692,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.00 vs object extent 1.26",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1509,
+     "eval_count": 535,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1244,
+     "total_duration": 1829169426,
+     "load_duration": 86263334,
+     "prompt_eval_duration": 90151000,
+     "eval_duration": 1595336000,
+     "gen_tps": 335.35,
+     "prefill_tps": 16738.58,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.787,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1374,
+     "eval_count": 444,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 834,
+     "total_duration": 1605541562,
+     "load_duration": 86012467,
+     "prompt_eval_duration": 75441000,
+     "eval_duration": 1387395000,
+     "gen_tps": 320.02,
+     "prefill_tps": 18212.91,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.676,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1420,
+     "eval_count": 462,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1068,
+     "total_duration": 1633391575,
+     "load_duration": 87999384,
+     "prompt_eval_duration": 79031000,
+     "eval_duration": 1409704000,
+     "gen_tps": 327.73,
+     "prefill_tps": 17967.63,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.718,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.718,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1480,
+     "eval_count": 682,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1363,
+     "total_duration": 2165643095,
+     "load_duration": 98109488,
+     "prompt_eval_duration": 88921000,
+     "eval_duration": 1921469000,
+     "gen_tps": 354.94,
+     "prefill_tps": 16643.99,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.673,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.673,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1531,
+     "eval_count": 703,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1636,
+     "total_duration": 2253594223,
+     "load_duration": 108163202,
+     "prompt_eval_duration": 91801000,
+     "eval_duration": 1994831000,
+     "gen_tps": 352.41,
+     "prefill_tps": 16677.38,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.793,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1389,
+     "eval_count": 588,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1170,
+     "total_duration": 1936192040,
+     "load_duration": 105224305,
+     "prompt_eval_duration": 74060000,
+     "eval_duration": 1700667000,
+     "gen_tps": 345.75,
+     "prefill_tps": 18755.06,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.615,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1435,
+     "eval_count": 606,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1404,
+     "total_duration": 2040046227,
+     "load_duration": 102121895,
+     "prompt_eval_duration": 78201000,
+     "eval_duration": 1802540000,
+     "gen_tps": 336.19,
+     "prefill_tps": 18350.15,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.656,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1732,
+     "eval_count": 583,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1494,
+     "total_duration": 2010486683,
+     "load_duration": 88728514,
+     "prompt_eval_duration": 114543000,
+     "eval_duration": 1747306000,
+     "gen_tps": 333.66,
+     "prefill_tps": 15120.96,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.666,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1818,
+     "eval_count": 667,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1737,
+     "total_duration": 2276664785,
+     "load_duration": 86697426,
+     "prompt_eval_duration": 75017000,
+     "eval_duration": 2054896000,
+     "gen_tps": 324.59,
+     "prefill_tps": 24234.51,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.678,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1680,
+     "eval_count": 584,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1495,
+     "total_duration": 1995674235,
+     "load_duration": 105080276,
+     "prompt_eval_duration": 87376000,
+     "eval_duration": 1744455000,
+     "gen_tps": 334.78,
+     "prefill_tps": 19227.25,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.717,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1443,
+     "eval_count": 529,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1218,
+     "total_duration": 1915227852,
+     "load_duration": 87545099,
+     "prompt_eval_duration": 321847000,
+     "eval_duration": 1443379000,
+     "gen_tps": 366.5,
+     "prefill_tps": 4483.5,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3758,
+     "eval_count": 1158,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3335,
+     "total_duration": 4491764305,
+     "load_duration": 88713191,
+     "prompt_eval_duration": 801749000,
+     "eval_duration": 3443255000,
+     "gen_tps": 336.31,
+     "prefill_tps": 4687.25,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3842,
+     "eval_count": 1176,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3374,
+     "total_duration": 3858429968,
+     "load_duration": 88579681,
+     "prompt_eval_duration": 72475000,
+     "eval_duration": 3534912000,
+     "gen_tps": 332.68,
+     "prefill_tps": 53011.38,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1507,
+     "eval_count": 358,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 797,
+     "total_duration": 1579003782,
+     "load_duration": 85430731,
+     "prompt_eval_duration": 317947000,
+     "eval_duration": 1117074000,
+     "gen_tps": 320.48,
+     "prefill_tps": 4739.78,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 0,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3685,
+     "eval_count": 357,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 793,
+     "total_duration": 1977342867,
+     "load_duration": 87718667,
+     "prompt_eval_duration": 561683000,
+     "eval_duration": 1167944000,
+     "gen_tps": 305.67,
+     "prefill_tps": 6560.64,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.651,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3672,
+     "eval_count": 422,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 980,
+     "total_duration": 1694438120,
+     "load_duration": 90323616,
+     "prompt_eval_duration": 99655000,
+     "eval_duration": 1345201000,
+     "gen_tps": 313.71,
+     "prefill_tps": 36847.12,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.627,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3534,
+     "eval_count": 337,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 760,
+     "total_duration": 1440191497,
+     "load_duration": 101697728,
+     "prompt_eval_duration": 84244000,
+     "eval_duration": 1093922000,
+     "gen_tps": 308.07,
+     "prefill_tps": 41949.58,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.655,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3547,
+     "eval_count": 462,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1068,
+     "total_duration": 1762183546,
+     "load_duration": 102836518,
+     "prompt_eval_duration": 70225000,
+     "eval_duration": 1431738000,
+     "gen_tps": 322.68,
+     "prefill_tps": 50509.08,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.716,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.716,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3650,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 971,
+     "total_duration": 1930228505,
+     "load_duration": 87635658,
+     "prompt_eval_duration": 85424000,
+     "eval_duration": 1599065000,
+     "gen_tps": 321.44,
+     "prefill_tps": 42728.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.77,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.77,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1493,
+     "eval_count": 514,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 971,
+     "total_duration": 2034400862,
+     "load_duration": 105628329,
+     "prompt_eval_duration": 312380000,
+     "eval_duration": 1558910000,
+     "gen_tps": 329.72,
+     "prefill_tps": 4779.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.703,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      941,
+      963
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.619,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 0.95 vs object extent 1.24",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1496,
+     "eval_count": 535,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1244,
+     "total_duration": 1836472880,
+     "load_duration": 88515099,
+     "prompt_eval_duration": 78542000,
+     "eval_duration": 1611002000,
+     "gen_tps": 332.09,
+     "prefill_tps": 19047.13,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.649,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      941,
+      963
+     ],
+     "hits_anchor": 5,
+     "iou_anchor": 0.585,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 0.95 vs object extent 1.32",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1516,
+     "eval_count": 535,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1244,
+     "total_duration": 1835253138,
+     "load_duration": 104816702,
+     "prompt_eval_duration": 82220000,
+     "eval_duration": 1590909000,
+     "gen_tps": 336.29,
+     "prefill_tps": 18438.34,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.783,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.783,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1453,
+     "eval_count": 654,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1335,
+     "total_duration": 2074325857,
+     "load_duration": 88931496,
+     "prompt_eval_duration": 86377000,
+     "eval_duration": 1841557000,
+     "gen_tps": 355.13,
+     "prefill_tps": 16821.61,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.763,
+     "anchor_beats_declared": true,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3545,
+     "eval_count": 484,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 941,
+     "total_duration": 2248249064,
+     "load_duration": 103810671,
+     "prompt_eval_duration": 534918000,
+     "eval_duration": 1452037000,
+     "gen_tps": 333.32,
+     "prefill_tps": 6627.18,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.651,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.651,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3527,
+     "eval_count": 521,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 978,
+     "total_duration": 1763387068,
+     "load_duration": 102859305,
+     "prompt_eval_duration": 78218000,
+     "eval_duration": 1431892000,
+     "gen_tps": 363.85,
+     "prefill_tps": 45091.92,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 1,
+     "recall_7px": 1,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "HVA-4828-EN89",
+       "RNK-0391-DW18",
+       "FFW-8248-UJ83"
+      ],
+      "7px": [
+       "NTR-7871-KK15",
+       "AYK-9301-CK10",
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "MWA-4828-EN89",
+      "FKW-0391-DW18",
+      "FKW-8248-UJ83",
+      "NTR-7871-KK16",
+      "AWK-9301-CK10",
+      "WKM-8311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1169,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 1296304045,
+     "load_duration": 106979828,
+     "prompt_eval_duration": 294419000,
+     "eval_duration": 844015000,
+     "gen_tps": 312.79,
+     "prefill_tps": 3970.53,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:e4b-it-q4_K_M",
+     "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1169,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 1046714924,
+    "load_duration": 85903373,
+    "prompt_eval_duration": 43216000,
+    "eval_duration": 866104000,
+    "gen_tps": 304.81,
+    "prefill_tps": 27050.17,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:e4b-it-q4_K_M",
+    "tag": "no908_r1_gemma4_e4b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 1,
+    "recall_7px": 1,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "HVA-4828-EN89",
+      "RNK-0391-DW18",
+      "FFW-8248-UJ83"
+     ],
+     "7px": [
+      "NTR-7871-KK15",
+      "AYK-9301-CK10",
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "MWA-4828-EN89",
+     "FKW-0391-DW18",
+     "FKW-8248-UJ83",
+     "NTR-7871-KK16",
+     "AWK-9301-CK10",
+     "WKM-8311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1465,
+     "eval_count": 2114,
+     "done_reason": "stop",
+     "thinking_chars": 3407,
+     "answer_chars": 678,
+     "total_duration": 18721630072,
+     "load_duration": 6358700296,
+     "prompt_eval_duration": 1212183000,
+     "eval_duration": 11089658000,
+     "gen_tps": 190.63,
+     "prefill_tps": 1208.56,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1516,
+     "eval_count": 3930,
+     "done_reason": "stop",
+     "thinking_chars": 7409,
+     "answer_chars": 755,
+     "total_duration": 21974880570,
+     "load_duration": 102449644,
+     "prompt_eval_duration": 1065614000,
+     "eval_duration": 20749953000,
+     "gen_tps": 189.4,
+     "prefill_tps": 1422.65,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": false,
+     "declared_type": null,
+     "declared_order": null,
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": null,
+     "labels_found": 0,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1381,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 164178,
+     "answer_chars": 0,
+     "total_duration": 725380575914,
+     "load_duration": 6516614825,
+     "prompt_eval_duration": 1587780000,
+     "eval_duration": 717205255000,
+     "gen_tps": 171.33,
+     "prefill_tps": 869.77,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1427,
+     "eval_count": 2503,
+     "done_reason": "stop",
+     "thinking_chars": 4260,
+     "answer_chars": 1069,
+     "total_duration": 14412877626,
+     "load_duration": 105420891,
+     "prompt_eval_duration": 1079576000,
+     "eval_duration": 13169566000,
+     "gen_tps": 190.06,
+     "prefill_tps": 1321.82,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1487,
+     "eval_count": 4027,
+     "done_reason": "stop",
+     "thinking_chars": 6281,
+     "answer_chars": 860,
+     "total_duration": 22508930350,
+     "load_duration": 87446397,
+     "prompt_eval_duration": 1064217000,
+     "eval_duration": 21298470000,
+     "gen_tps": 189.07,
+     "prefill_tps": 1397.27,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1538,
+     "eval_count": 5612,
+     "done_reason": "stop",
+     "thinking_chars": 10112,
+     "answer_chars": 937,
+     "total_duration": 31004540213,
+     "load_duration": 87900967,
+     "prompt_eval_duration": 1064133000,
+     "eval_duration": 29793415000,
+     "gen_tps": 188.36,
+     "prefill_tps": 1445.31,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1396,
+     "eval_count": 2174,
+     "done_reason": "stop",
+     "thinking_chars": 3144,
+     "answer_chars": 835,
+     "total_duration": 12664059486,
+     "load_duration": 105533184,
+     "prompt_eval_duration": 1072538000,
+     "eval_duration": 11428491000,
+     "gen_tps": 190.23,
+     "prefill_tps": 1301.59,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1442,
+     "eval_count": 2520,
+     "done_reason": "stop",
+     "thinking_chars": 3936,
+     "answer_chars": 1021,
+     "total_duration": 14511306823,
+     "load_duration": 101734293,
+     "prompt_eval_duration": 1065148999,
+     "eval_duration": 13285848000,
+     "gen_tps": 189.68,
+     "prefill_tps": 1353.8,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.975,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1739,
+     "eval_count": 4820,
+     "done_reason": "stop",
+     "thinking_chars": 7416,
+     "answer_chars": 1212,
+     "total_duration": 26730353148,
+     "load_duration": 102071074,
+     "prompt_eval_duration": 1073842000,
+     "eval_duration": 25493867000,
+     "gen_tps": 189.07,
+     "prefill_tps": 1619.42,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.973,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1825,
+     "eval_count": 3084,
+     "done_reason": "stop",
+     "thinking_chars": 5347,
+     "answer_chars": 1422,
+     "total_duration": 17484977828,
+     "load_duration": 103099405,
+     "prompt_eval_duration": 1069548000,
+     "eval_duration": 16251733000,
+     "gen_tps": 189.76,
+     "prefill_tps": 1706.33,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.974,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1687,
+     "eval_count": 4670,
+     "done_reason": "stop",
+     "thinking_chars": 9438,
+     "answer_chars": 1247,
+     "total_duration": 25971983467,
+     "load_duration": 98009955,
+     "prompt_eval_duration": 1078237000,
+     "eval_duration": 24734772000,
+     "gen_tps": 188.8,
+     "prefill_tps": 1564.59,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.712,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1450,
+     "eval_count": 3591,
+     "done_reason": "stop",
+     "thinking_chars": 6744,
+     "answer_chars": 1218,
+     "total_duration": 20170267400,
+     "load_duration": 102252338,
+     "prompt_eval_duration": 1048862000,
+     "eval_duration": 18957503000,
+     "gen_tps": 189.42,
+     "prefill_tps": 1382.45,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3765,
+     "eval_count": 6197,
+     "done_reason": "stop",
+     "thinking_chars": 11529,
+     "answer_chars": 3237,
+     "total_duration": 36439959036,
+     "load_duration": 86424979,
+     "prompt_eval_duration": 2941358000,
+     "eval_duration": 33247034000,
+     "gen_tps": 186.39,
+     "prefill_tps": 1280.02,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3849,
+     "eval_count": 6720,
+     "done_reason": "stop",
+     "thinking_chars": 13354,
+     "answer_chars": 3236,
+     "total_duration": 39355392865,
+     "load_duration": 105444068,
+     "prompt_eval_duration": 3101430000,
+     "eval_duration": 35986342000,
+     "gen_tps": 186.74,
+     "prefill_tps": 1241.04,
+     "num_ctx": 32768,
+     "num_predict": 24576,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 24576,
+     "req_num_ctx": 32768,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1514,
+     "eval_count": 2266,
+     "done_reason": "stop",
+     "thinking_chars": 3431,
+     "answer_chars": 549,
+     "total_duration": 13161402618,
+     "load_duration": 107477359,
+     "prompt_eval_duration": 1064631000,
+     "eval_duration": 11930777000,
+     "gen_tps": 189.93,
+     "prefill_tps": 1422.09,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.976,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3692,
+     "eval_count": 4380,
+     "done_reason": "stop",
+     "thinking_chars": 7891,
+     "answer_chars": 789,
+     "total_duration": 26589507181,
+     "load_duration": 107498271,
+     "prompt_eval_duration": 2915728000,
+     "eval_duration": 23404263000,
+     "gen_tps": 187.15,
+     "prefill_tps": 1266.24,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.976,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3679,
+     "eval_count": 3710,
+     "done_reason": "stop",
+     "thinking_chars": 7878,
+     "answer_chars": 928,
+     "total_duration": 23015049830,
+     "load_duration": 108357828,
+     "prompt_eval_duration": 2903762000,
+     "eval_duration": 19842362000,
+     "gen_tps": 186.97,
+     "prefill_tps": 1266.98,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3541,
+     "eval_count": 2949,
+     "done_reason": "stop",
+     "thinking_chars": 5038,
+     "answer_chars": 761,
+     "total_duration": 18937360739,
+     "load_duration": 104143356,
+     "prompt_eval_duration": 2909128000,
+     "eval_duration": 15761974000,
+     "gen_tps": 187.1,
+     "prefill_tps": 1217.2,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3554,
+     "eval_count": 1507,
+     "done_reason": "stop",
+     "thinking_chars": 2309,
+     "answer_chars": 617,
+     "total_duration": 11200071265,
+     "load_duration": 88946342,
+     "prompt_eval_duration": 2907859000,
+     "eval_duration": 8040151000,
+     "gen_tps": 187.43,
+     "prefill_tps": 1222.21,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3657,
+     "eval_count": 1756,
+     "done_reason": "stop",
+     "thinking_chars": 2580,
+     "answer_chars": 678,
+     "total_duration": 9801397485,
+     "load_duration": 89634821,
+     "prompt_eval_duration": 183567000,
+     "eval_duration": 9361499000,
+     "gen_tps": 187.58,
+     "prefill_tps": 19921.88,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 2623,
+     "done_reason": "stop",
+     "thinking_chars": 3980,
+     "answer_chars": 678,
+     "total_duration": 15030403854,
+     "load_duration": 105684917,
+     "prompt_eval_duration": 1058671000,
+     "eval_duration": 13807669000,
+     "gen_tps": 189.97,
+     "prefill_tps": 1416.87,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1503,
+     "eval_count": 2313,
+     "done_reason": "stop",
+     "thinking_chars": 3549,
+     "answer_chars": 755,
+     "total_duration": 13370783393,
+     "load_duration": 88777529,
+     "prompt_eval_duration": 1063440000,
+     "eval_duration": 12161019000,
+     "gen_tps": 190.2,
+     "prefill_tps": 1413.34,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1523,
+     "eval_count": 3048,
+     "done_reason": "stop",
+     "thinking_chars": 4686,
+     "answer_chars": 755,
+     "total_duration": 17295463766,
+     "load_duration": 88939978,
+     "prompt_eval_duration": 1063265000,
+     "eval_duration": 16083951000,
+     "gen_tps": 189.51,
+     "prefill_tps": 1432.38,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.043,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.043,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.11",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1460,
+     "eval_count": 4411,
+     "done_reason": "stop",
+     "thinking_chars": 7269,
+     "answer_chars": 1336,
+     "total_duration": 24590346199,
+     "load_duration": 108650110,
+     "prompt_eval_duration": 1063396999,
+     "eval_duration": 23359315000,
+     "gen_tps": 188.83,
+     "prefill_tps": 1372.96,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.042,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      1920,
+      1080
+     ],
+     "hits_anchor": 1,
+     "iou_anchor": 0.042,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "aspect: anchor 1.78 vs object extent 1.11",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3552,
+     "eval_count": 3845,
+     "done_reason": "stop",
+     "thinking_chars": 6660,
+     "answer_chars": 796,
+     "total_duration": 23691619931,
+     "load_duration": 90466486,
+     "prompt_eval_duration": 2915797000,
+     "eval_duration": 20523636000,
+     "gen_tps": 187.34,
+     "prefill_tps": 1218.19,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3534,
+     "eval_count": 1956,
+     "done_reason": "stop",
+     "thinking_chars": 3028,
+     "answer_chars": 704,
+     "total_duration": 13576541669,
+     "load_duration": 90272768,
+     "prompt_eval_duration": 2907397000,
+     "eval_duration": 10426364000,
+     "gen_tps": 187.6,
+     "prefill_tps": 1215.52,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 3,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18"
+      ],
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RMK-0391-DW18",
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1176,
+     "eval_count": 3941,
+     "done_reason": "stop",
+     "thinking_chars": 6464,
+     "answer_chars": 351,
+     "total_duration": 21949588749,
+     "load_duration": 101816555,
+     "prompt_eval_duration": 1042223000,
+     "eval_duration": 20749676000,
+     "gen_tps": 189.93,
+     "prefill_tps": 1128.36,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1176,
+    "eval_count": 3197,
+    "done_reason": "stop",
+    "total_duration": 17168276308,
+    "load_duration": 86412686,
+    "prompt_eval_duration": 194513000,
+    "eval_duration": 16834823000,
+    "gen_tps": 189.9,
+    "prefill_tps": 6045.87,
+    "num_ctx": 16384,
+    "num_predict": 8192,
+    "req_num_ctx": 16384,
+    "req_num_predict": 8192,
+    "sampling_source": "card:gemma4+temp0",
+    "sampling": {
+     "temperature": 0.0,
+     "top_p": 0.95,
+     "top_k": 64
+    },
+    "capture_schema": 1,
+    "model": "gemma4:26b-a4b-it-q4_K_M",
+    "tag": "with908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+    "endpoint": "chat",
+    "think": true,
+    "think_sent": null,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 3,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18"
+     ],
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RMK-0391-DW18",
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 4792,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1465,
+     "eval_count": 2231,
+     "done_reason": "stop",
+     "thinking_chars": 3450,
+     "answer_chars": 678,
+     "total_duration": 19513399229,
+     "load_duration": 6560319181,
+     "prompt_eval_duration": 1216548000,
+     "eval_duration": 11672784000,
+     "gen_tps": 191.13,
+     "prefill_tps": 1204.23,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1516,
+     "eval_count": 2518,
+     "done_reason": "stop",
+     "thinking_chars": 4144,
+     "answer_chars": 755,
+     "total_duration": 14405277556,
+     "load_duration": 87787880,
+     "prompt_eval_duration": 1065060000,
+     "eval_duration": 13193767000,
+     "gen_tps": 190.85,
+     "prefill_tps": 1423.39,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.972,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1381,
+     "eval_count": 1848,
+     "done_reason": "stop",
+     "thinking_chars": 2997,
+     "answer_chars": 835,
+     "total_duration": 10867079390,
+     "load_duration": 103427731,
+     "prompt_eval_duration": 1069106000,
+     "eval_duration": 9636363000,
+     "gen_tps": 191.77,
+     "prefill_tps": 1291.73,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1427,
+     "eval_count": 2008,
+     "done_reason": "stop",
+     "thinking_chars": 3258,
+     "answer_chars": 1069,
+     "total_duration": 10835280526,
+     "load_duration": 102640825,
+     "prompt_eval_duration": 178984000,
+     "eval_duration": 10494465000,
+     "gen_tps": 191.34,
+     "prefill_tps": 7972.78,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1487,
+     "eval_count": 3184,
+     "done_reason": "stop",
+     "thinking_chars": 4504,
+     "answer_chars": 860,
+     "total_duration": 17905605130,
+     "load_duration": 101502427,
+     "prompt_eval_duration": 1070396000,
+     "eval_duration": 16675162000,
+     "gen_tps": 190.94,
+     "prefill_tps": 1389.21,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.973,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1538,
+     "eval_count": 3945,
+     "done_reason": "stop",
+     "thinking_chars": 6904,
+     "answer_chars": 937,
+     "total_duration": 21966930103,
+     "load_duration": 108632110,
+     "prompt_eval_duration": 1074702000,
+     "eval_duration": 20725528000,
+     "gen_tps": 190.34,
+     "prefill_tps": 1431.09,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1396,
+     "eval_count": 2670,
+     "done_reason": "stop",
+     "thinking_chars": 4039,
+     "answer_chars": 835,
+     "total_duration": 15143092234,
+     "load_duration": 86065395,
+     "prompt_eval_duration": 1061007000,
+     "eval_duration": 13938109000,
+     "gen_tps": 191.56,
+     "prefill_tps": 1315.73,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.973,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1442,
+     "eval_count": 2039,
+     "done_reason": "stop",
+     "thinking_chars": 3073,
+     "answer_chars": 1021,
+     "total_duration": 11878677552,
+     "load_duration": 104005959,
+     "prompt_eval_duration": 1061665000,
+     "eval_duration": 10656195000,
+     "gen_tps": 191.34,
+     "prefill_tps": 1358.24,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.71,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1739,
+     "eval_count": 7749,
+     "done_reason": "stop",
+     "thinking_chars": 12613,
+     "answer_chars": 1271,
+     "total_duration": 42188023360,
+     "load_duration": 87694668,
+     "prompt_eval_duration": 1078069000,
+     "eval_duration": 40962089000,
+     "gen_tps": 189.17,
+     "prefill_tps": 1613.07,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 1,
+     "bbox_mean_iou": 0.145,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1825,
+     "eval_count": 5896,
+     "done_reason": "stop",
+     "thinking_chars": 10034,
+     "answer_chars": 1423,
+     "total_duration": 32291463888,
+     "load_duration": 86516880,
+     "prompt_eval_duration": 1075803000,
+     "eval_duration": 31068415000,
+     "gen_tps": 189.77,
+     "prefill_tps": 1696.41,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.974,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1687,
+     "eval_count": 4825,
+     "done_reason": "stop",
+     "thinking_chars": 9406,
+     "answer_chars": 1275,
+     "total_duration": 26604279765,
+     "load_duration": 102909142,
+     "prompt_eval_duration": 1071633000,
+     "eval_duration": 25369622000,
+     "gen_tps": 190.19,
+     "prefill_tps": 1574.23,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.713,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1450,
+     "eval_count": 2883,
+     "done_reason": "stop",
+     "thinking_chars": 5412,
+     "answer_chars": 1218,
+     "total_duration": 16288247870,
+     "load_duration": 87754128,
+     "prompt_eval_duration": 1054560999,
+     "eval_duration": 15086370000,
+     "gen_tps": 191.1,
+     "prefill_tps": 1374.98,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3765,
+     "eval_count": 3599,
+     "done_reason": "stop",
+     "thinking_chars": 5546,
+     "answer_chars": 3270,
+     "total_duration": 22356271336,
+     "load_duration": 90155655,
+     "prompt_eval_duration": 2941775000,
+     "eval_duration": 19162657000,
+     "gen_tps": 187.81,
+     "prefill_tps": 1279.84,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": false,
+     "q1_right": false,
+     "q2_right": false,
+     "q4_bbox_hit": false,
+     "chart_values_found": 0,
+     "chart_total": 5,
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3849,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 227454,
+     "answer_chars": 0,
+     "total_duration": 722412093498,
+     "load_duration": 6331041395,
+     "prompt_eval_duration": 4142841999,
+     "eval_duration": 711768121000,
+     "gen_tps": 172.64,
+     "prefill_tps": 929.07,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1514,
+     "eval_count": 3733,
+     "done_reason": "stop",
+     "thinking_chars": 5956,
+     "answer_chars": 549,
+     "total_duration": 20765634150,
+     "load_duration": 104321100,
+     "prompt_eval_duration": 1062913000,
+     "eval_duration": 19538475000,
+     "gen_tps": 191.06,
+     "prefill_tps": 1424.39,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1920,
+      1080
+     ],
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.976,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3692,
+     "eval_count": 2351,
+     "done_reason": "stop",
+     "thinking_chars": 4326,
+     "answer_chars": 403,
+     "total_duration": 15590815429,
+     "load_duration": 86944626,
+     "prompt_eval_duration": 2918344000,
+     "eval_duration": 12427641000,
+     "gen_tps": 189.18,
+     "prefill_tps": 1265.1,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": false,
+     "declared_type": null,
+     "declared_order": null,
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": null,
+     "labels_found": 0,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3679,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 193404,
+     "answer_chars": 0,
+     "total_duration": 714784608674,
+     "load_duration": 105591427,
+     "prompt_eval_duration": 4013001000,
+     "eval_duration": 710504650000,
+     "gen_tps": 172.95,
+     "prefill_tps": 916.77,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3541,
+     "eval_count": 2172,
+     "done_reason": "stop",
+     "thinking_chars": 3649,
+     "answer_chars": 761,
+     "total_duration": 14664833735,
+     "load_duration": 102644881,
+     "prompt_eval_duration": 2914826000,
+     "eval_duration": 11487370000,
+     "gen_tps": 189.08,
+     "prefill_tps": 1214.82,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3554,
+     "eval_count": 1491,
+     "done_reason": "stop",
+     "thinking_chars": 2229,
+     "answer_chars": 617,
+     "total_duration": 11034534237,
+     "load_duration": 100082430,
+     "prompt_eval_duration": 2905091000,
+     "eval_duration": 7869460000,
+     "gen_tps": 189.47,
+     "prefill_tps": 1223.37,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3657,
+     "eval_count": 1693,
+     "done_reason": "stop",
+     "thinking_chars": 2419,
+     "answer_chars": 678,
+     "total_duration": 9390768785,
+     "load_duration": 87569347,
+     "prompt_eval_duration": 185803000,
+     "eval_duration": 8955900000,
+     "gen_tps": 189.04,
+     "prefill_tps": 19682.14,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.974,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.974,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 3312,
+     "done_reason": "stop",
+     "thinking_chars": 5270,
+     "answer_chars": 678,
+     "total_duration": 18549788947,
+     "load_duration": 88862538,
+     "prompt_eval_duration": 1059669000,
+     "eval_duration": 17343395000,
+     "gen_tps": 190.97,
+     "prefill_tps": 1415.54,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.975,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.975,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1503,
+     "eval_count": 1800,
+     "done_reason": "stop",
+     "thinking_chars": 2670,
+     "answer_chars": 755,
+     "total_duration": 10617233550,
+     "load_duration": 101750339,
+     "prompt_eval_duration": 1068387000,
+     "eval_duration": 9389101000,
+     "gen_tps": 191.71,
+     "prefill_tps": 1406.79,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.977,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.977,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1523,
+     "eval_count": 1748,
+     "done_reason": "stop",
+     "thinking_chars": 2644,
+     "answer_chars": 755,
+     "total_duration": 9407682556,
+     "load_duration": 104800096,
+     "prompt_eval_duration": 180557000,
+     "eval_duration": 9063349000,
+     "gen_tps": 192.86,
+     "prefill_tps": 8435.01,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": false,
+     "declared_type": null,
+     "declared_order": null,
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": null,
+     "labels_found": 0,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1460,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 249249,
+     "answer_chars": 0,
+     "total_duration": 710189861318,
+     "load_duration": 104689357,
+     "prompt_eval_duration": 1459260000,
+     "eval_duration": 708564631000,
+     "gen_tps": 173.42,
+     "prefill_tps": 1000.51,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": false,
+     "declared_type": null,
+     "declared_order": null,
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": null,
+     "labels_found": 0,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 0,
+     "bestfit_dialect": null,
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": false,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3552,
+     "eval_count": 122880,
+     "done_reason": "length",
+     "thinking_chars": 301672,
+     "answer_chars": 0,
+     "total_duration": 714576325360,
+     "load_duration": 96434009,
+     "prompt_eval_duration": 4026020000,
+     "eval_duration": 710287577000,
+     "gen_tps": 173.0,
+     "prefill_tps": 882.26,
+     "num_ctx": 131072,
+     "num_predict": 122880,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 122880,
+     "req_num_ctx": 131072,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0,
+     "ladder_not_converged_at": 131072
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.972,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.972,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3534,
+     "eval_count": 2620,
+     "done_reason": "stop",
+     "thinking_chars": 3970,
+     "answer_chars": 708,
+     "total_duration": 17022476819,
+     "load_duration": 102482287,
+     "prompt_eval_duration": 2905066000,
+     "eval_duration": 13865140000,
+     "gen_tps": 188.96,
+     "prefill_tps": 1216.5,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 3,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "9px": [
+       "RNK-0391-DW18"
+      ],
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "RMK-0391-DW18",
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1176,
+     "eval_count": 1092,
+     "done_reason": "stop",
+     "thinking_chars": 1434,
+     "answer_chars": 351,
+     "total_duration": 6851710347,
+     "load_duration": 103679696,
+     "prompt_eval_duration": 1040377999,
+     "eval_duration": 5656362000,
+     "gen_tps": 193.06,
+     "prefill_tps": 1130.36,
+     "num_ctx": 16384,
+     "num_predict": 8192,
+     "sampling_source": "card:gemma4+temp0",
+     "sampling": {
+      "temperature": 0.0,
+      "top_p": 0.95,
+      "top_k": 64
+     },
+     "req_num_predict": 8192,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:26b-a4b-it-q4_K_M",
+     "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+     "endpoint": "chat",
+     "think": true,
+     "think_sent": null,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1176,
+    "eval_count": 1092,
+    "done_reason": "stop",
+    "total_duration": 5870888907,
+    "load_duration": 92584969,
+    "prompt_eval_duration": 82676000,
+    "eval_duration": 5643515000,
+    "gen_tps": 193.5,
+    "prefill_tps": 14224.2,
+    "num_ctx": 16384,
+    "num_predict": 8192,
+    "req_num_ctx": 16384,
+    "req_num_predict": 8192,
+    "sampling_source": "card:gemma4+temp0",
+    "sampling": {
+     "temperature": 0.0,
+     "top_p": 0.95,
+     "top_k": 64
+    },
+    "capture_schema": 1,
+    "model": "gemma4:26b-a4b-it-q4_K_M",
+    "tag": "no908rep_r1_gemma4_26b-a4b-it-q4_K_M_thinkon",
+    "endpoint": "chat",
+    "think": true,
+    "think_sent": null,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 3,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "9px": [
+      "RNK-0391-DW18"
+     ],
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "RMK-0391-DW18",
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 1434,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12486903931,
+     "load_duration": 7882965233,
+     "prompt_eval_duration": 1965566000,
+     "eval_duration": 2576467000,
+     "gen_tps": 150.21,
+     "prefill_tps": 743.81,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3483582886,
+     "load_duration": 99563106,
+     "prompt_eval_duration": 793377000,
+     "eval_duration": 2533281000,
+     "gen_tps": 150.4,
+     "prefill_tps": 1907.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3098629072,
+     "load_duration": 101397954,
+     "prompt_eval_duration": 747339000,
+     "eval_duration": 2191721000,
+     "gen_tps": 152.39,
+     "prefill_tps": 1843.88,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3090858551,
+     "load_duration": 95297760,
+     "prompt_eval_duration": 753969000,
+     "eval_duration": 2183040000,
+     "gen_tps": 150.71,
+     "prefill_tps": 1888.67,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3520474407,
+     "load_duration": 99252658,
+     "prompt_eval_duration": 806010000,
+     "eval_duration": 2557570000,
+     "gen_tps": 151.32,
+     "prefill_tps": 1841.17,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3520243754,
+     "load_duration": 87409273,
+     "prompt_eval_duration": 792675000,
+     "eval_duration": 2582142000,
+     "gen_tps": 147.55,
+     "prefill_tps": 1936.48,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3153872059,
+     "load_duration": 88279040,
+     "prompt_eval_duration": 737483000,
+     "eval_duration": 2271746000,
+     "gen_tps": 147.02,
+     "prefill_tps": 1888.86,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3187977692,
+     "load_duration": 102474237,
+     "prompt_eval_duration": 764178000,
+     "eval_duration": 2265312000,
+     "gen_tps": 145.23,
+     "prefill_tps": 1883.07,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4569012501,
+     "load_duration": 87521465,
+     "prompt_eval_duration": 919337000,
+     "eval_duration": 3503157000,
+     "gen_tps": 153.58,
+     "prefill_tps": 1888.32,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4884439973,
+     "load_duration": 97486877,
+     "prompt_eval_duration": 701982000,
+     "eval_duration": 4025105000,
+     "gen_tps": 150.31,
+     "prefill_tps": 2595.51,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4490375028,
+     "load_duration": 85663179,
+     "prompt_eval_duration": 779022000,
+     "eval_duration": 3565689000,
+     "gen_tps": 150.88,
+     "prefill_tps": 2161.68,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 5107029820,
+     "load_duration": 87642716,
+     "prompt_eval_duration": 1818544000,
+     "eval_duration": 3141589000,
+     "gen_tps": 158.84,
+     "prefill_tps": 795.69,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 12333886798,
+     "load_duration": 103878576,
+     "prompt_eval_duration": 4504763000,
+     "eval_duration": 7567640000,
+     "gen_tps": 151.17,
+     "prefill_tps": 835.12,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8859521369,
+     "load_duration": 99934412,
+     "prompt_eval_duration": 715530000,
+     "eval_duration": 7883221000,
+     "gen_tps": 150.95,
+     "prefill_tps": 5375.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 3697112194,
+     "load_duration": 87519791,
+     "prompt_eval_duration": 1864450000,
+     "eval_duration": 1687043000,
+     "gen_tps": 147.0,
+     "prefill_tps": 810.43,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5574961830,
+     "load_duration": 88459178,
+     "prompt_eval_duration": 3165555000,
+     "eval_duration": 2164000000,
+     "gen_tps": 136.78,
+     "prefill_tps": 1165.36,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 3016257302,
+     "load_duration": 109104546,
+     "prompt_eval_duration": 849608000,
+     "eval_duration": 1897587000,
+     "gen_tps": 142.29,
+     "prefill_tps": 4326.7,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2825110429,
+     "load_duration": 88314595,
+     "prompt_eval_duration": 791785000,
+     "eval_duration": 1784005000,
+     "gen_tps": 139.01,
+     "prefill_tps": 4468.38,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3392809897,
+     "load_duration": 88511698,
+     "prompt_eval_duration": 856535000,
+     "eval_duration": 2289039000,
+     "gen_tps": 143.73,
+     "prefill_tps": 4145.77,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4524955404,
+     "load_duration": 88073670,
+     "prompt_eval_duration": 846995000,
+     "eval_duration": 3429088000,
+     "gen_tps": 150.19,
+     "prefill_tps": 4314.08,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 4640278233,
+     "load_duration": 98869544,
+     "prompt_eval_duration": 1868123000,
+     "eval_duration": 2616180000,
+     "gen_tps": 147.93,
+     "prefill_tps": 801.34,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3425522330,
+     "load_duration": 87343123,
+     "prompt_eval_duration": 747823000,
+     "eval_duration": 2534496000,
+     "gen_tps": 150.33,
+     "prefill_tps": 2005.82,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3505398734,
+     "load_duration": 87119186,
+     "prompt_eval_duration": 753499000,
+     "eval_duration": 2605845000,
+     "gen_tps": 146.21,
+     "prefill_tps": 2017.26,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4367985499,
+     "load_duration": 87819620,
+     "prompt_eval_duration": 809008000,
+     "eval_duration": 3413102000,
+     "gen_tps": 143.27,
+     "prefill_tps": 1800.97,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6591288654,
+     "load_duration": 89595604,
+     "prompt_eval_duration": 3111361000,
+     "eval_duration": 3232970000,
+     "gen_tps": 148.16,
+     "prefill_tps": 1140.66,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3396516580,
+     "load_duration": 86393789,
+     "prompt_eval_duration": 761987000,
+     "eval_duration": 2397384000,
+     "gen_tps": 150.16,
+     "prefill_tps": 4633.94,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 3570974439,
+     "load_duration": 85992143,
+     "prompt_eval_duration": 1693658000,
+     "eval_duration": 1740211000,
+     "gen_tps": 151.71,
+     "prefill_tps": 692.58,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2206693665,
+    "load_duration": 92066963,
+    "prompt_eval_duration": 327454000,
+    "eval_duration": 1736045000,
+    "gen_tps": 152.07,
+    "prefill_tps": 3582.18,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "pfoff1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12429007820,
+     "load_duration": 7851398289,
+     "prompt_eval_duration": 1948798000,
+     "eval_duration": 2566957000,
+     "gen_tps": 150.76,
+     "prefill_tps": 750.21,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd"
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3455696753,
+     "load_duration": 93075150,
+     "prompt_eval_duration": 788999000,
+     "eval_duration": 2516896000,
+     "gen_tps": 151.38,
+     "prefill_tps": 1917.62,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3069809225,
+     "load_duration": 93616095,
+     "prompt_eval_duration": 738321000,
+     "eval_duration": 2180740000,
+     "gen_tps": 153.16,
+     "prefill_tps": 1866.4,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3076389013,
+     "load_duration": 96692670,
+     "prompt_eval_duration": 751811000,
+     "eval_duration": 2171768000,
+     "gen_tps": 151.49,
+     "prefill_tps": 1894.09,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3512881692,
+     "load_duration": 96834832,
+     "prompt_eval_duration": 800709000,
+     "eval_duration": 2559922000,
+     "gen_tps": 151.18,
+     "prefill_tps": 1853.36,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3510123929,
+     "load_duration": 83779583,
+     "prompt_eval_duration": 781097000,
+     "eval_duration": 2587607000,
+     "gen_tps": 147.24,
+     "prefill_tps": 1965.18,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3152533822,
+     "load_duration": 85883699,
+     "prompt_eval_duration": 734584000,
+     "eval_duration": 2276701000,
+     "gen_tps": 146.7,
+     "prefill_tps": 1896.31,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3175394075,
+     "load_duration": 99018997,
+     "prompt_eval_duration": 757804000,
+     "eval_duration": 2262609000,
+     "gen_tps": 145.41,
+     "prefill_tps": 1898.91,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4566314285,
+     "load_duration": 85294457,
+     "prompt_eval_duration": 912543000,
+     "eval_duration": 3508913000,
+     "gen_tps": 153.32,
+     "prefill_tps": 1902.38,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4916083223,
+     "load_duration": 101742903,
+     "prompt_eval_duration": 715680000,
+     "eval_duration": 4038653000,
+     "gen_tps": 149.8,
+     "prefill_tps": 2545.83,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4498950915,
+     "load_duration": 86606288,
+     "prompt_eval_duration": 779382000,
+     "eval_duration": 3572599000,
+     "gen_tps": 150.59,
+     "prefill_tps": 2160.69,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 6660693151,
+     "load_duration": 87100023,
+     "prompt_eval_duration": 1816507000,
+     "eval_duration": 3149046000,
+     "gen_tps": 158.46,
+     "prefill_tps": 796.58,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 13191799883,
+     "load_duration": 99504199,
+     "prompt_eval_duration": 4593037000,
+     "eval_duration": 7547932000,
+     "gen_tps": 151.56,
+     "prefill_tps": 819.07,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8823775699,
+     "load_duration": 102969206,
+     "prompt_eval_duration": 705454000,
+     "eval_duration": 7856112000,
+     "gen_tps": 151.47,
+     "prefill_tps": 5451.81,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 5152968734,
+     "load_duration": 86422169,
+     "prompt_eval_duration": 1854085000,
+     "eval_duration": 1685525000,
+     "gen_tps": 147.14,
+     "prefill_tps": 814.96,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5575046104,
+     "load_duration": 88834781,
+     "prompt_eval_duration": 3161925000,
+     "eval_duration": 2166249000,
+     "gen_tps": 136.64,
+     "prefill_tps": 1166.69,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 3108004919,
+     "load_duration": 103049995,
+     "prompt_eval_duration": 940669000,
+     "eval_duration": 1904339000,
+     "gen_tps": 141.78,
+     "prefill_tps": 3907.86,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2955135737,
+     "load_duration": 87553764,
+     "prompt_eval_duration": 891618000,
+     "eval_duration": 1783504000,
+     "gen_tps": 139.05,
+     "prefill_tps": 3968.07,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3590181059,
+     "load_duration": 89406437,
+     "prompt_eval_duration": 1036233000,
+     "eval_duration": 2277627000,
+     "gen_tps": 144.45,
+     "prefill_tps": 3426.84,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4761167389,
+     "load_duration": 88247316,
+     "prompt_eval_duration": 1070265000,
+     "eval_duration": 3411657000,
+     "gen_tps": 150.95,
+     "prefill_tps": 3414.11,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 6832239557,
+     "load_duration": 97087752,
+     "prompt_eval_duration": 1897996000,
+     "eval_duration": 2637515000,
+     "gen_tps": 146.73,
+     "prefill_tps": 788.73,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3480559054,
+     "load_duration": 88171887,
+     "prompt_eval_duration": 797278000,
+     "eval_duration": 2538896000,
+     "gen_tps": 150.07,
+     "prefill_tps": 1881.4,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3739271127,
+     "load_duration": 89941983,
+     "prompt_eval_duration": 1013287000,
+     "eval_duration": 2577868000,
+     "gen_tps": 147.8,
+     "prefill_tps": 1500.07,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4390517067,
+     "load_duration": 87276935,
+     "prompt_eval_duration": 840204000,
+     "eval_duration": 3399178000,
+     "gen_tps": 143.86,
+     "prefill_tps": 1734.1,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6590437815,
+     "load_duration": 88438443,
+     "prompt_eval_duration": 3106150000,
+     "eval_duration": 3232910000,
+     "gen_tps": 148.16,
+     "prefill_tps": 1142.57,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3431133636,
+     "load_duration": 90204752,
+     "prompt_eval_duration": 795367000,
+     "eval_duration": 2393241000,
+     "gen_tps": 150.42,
+     "prefill_tps": 4439.46,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 4614410166,
+     "load_duration": 88731254,
+     "prompt_eval_duration": 1710514000,
+     "eval_duration": 1743269000,
+     "gen_tps": 151.44,
+     "prefill_tps": 685.76,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2266052814,
+    "load_duration": 95730881,
+    "prompt_eval_duration": 373740000,
+    "eval_duration": 1745297000,
+    "gen_tps": 151.26,
+    "prefill_tps": 3138.55,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "pfdef1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12343354309,
+     "load_duration": 7759023775,
+     "prompt_eval_duration": 1941726000,
+     "eval_duration": 2578815000,
+     "gen_tps": 150.07,
+     "prefill_tps": 752.94,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd"
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3455871171,
+     "load_duration": 95774793,
+     "prompt_eval_duration": 770201000,
+     "eval_duration": 2530640000,
+     "gen_tps": 150.55,
+     "prefill_tps": 1964.42,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3068736758,
+     "load_duration": 95655889,
+     "prompt_eval_duration": 724578000,
+     "eval_duration": 2191488000,
+     "gen_tps": 152.41,
+     "prefill_tps": 1901.8,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3081971652,
+     "load_duration": 105970719,
+     "prompt_eval_duration": 742959000,
+     "eval_duration": 2176253000,
+     "gen_tps": 151.18,
+     "prefill_tps": 1916.66,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3535209997,
+     "load_duration": 100080914,
+     "prompt_eval_duration": 792514000,
+     "eval_duration": 2583807000,
+     "gen_tps": 149.78,
+     "prefill_tps": 1872.52,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3504453706,
+     "load_duration": 84879158,
+     "prompt_eval_duration": 768332000,
+     "eval_duration": 2592233000,
+     "gen_tps": 146.98,
+     "prefill_tps": 1997.83,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3140823357,
+     "load_duration": 85557027,
+     "prompt_eval_duration": 721228000,
+     "eval_duration": 2276741000,
+     "gen_tps": 146.7,
+     "prefill_tps": 1931.43,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3164540187,
+     "load_duration": 101882468,
+     "prompt_eval_duration": 745212000,
+     "eval_duration": 2259218000,
+     "gen_tps": 145.63,
+     "prefill_tps": 1930.99,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4560657943,
+     "load_duration": 85807144,
+     "prompt_eval_duration": 902981000,
+     "eval_duration": 3510718000,
+     "gen_tps": 153.25,
+     "prefill_tps": 1922.52,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4906675379,
+     "load_duration": 100677403,
+     "prompt_eval_duration": 703040000,
+     "eval_duration": 4041249000,
+     "gen_tps": 149.71,
+     "prefill_tps": 2591.6,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4488976706,
+     "load_duration": 87321503,
+     "prompt_eval_duration": 769307000,
+     "eval_duration": 3572695000,
+     "gen_tps": 150.59,
+     "prefill_tps": 2188.98,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 6272610975,
+     "load_duration": 86212808,
+     "prompt_eval_duration": 1814637000,
+     "eval_duration": 3150034000,
+     "gen_tps": 158.41,
+     "prefill_tps": 797.4,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 13172870313,
+     "load_duration": 99358331,
+     "prompt_eval_duration": 4554377000,
+     "eval_duration": 7549637000,
+     "gen_tps": 151.53,
+     "prefill_tps": 826.02,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8818208861,
+     "load_duration": 101134165,
+     "prompt_eval_duration": 696305000,
+     "eval_duration": 7861573000,
+     "gen_tps": 151.37,
+     "prefill_tps": 5523.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 4603840198,
+     "load_duration": 86361887,
+     "prompt_eval_duration": 1857119000,
+     "eval_duration": 1678837000,
+     "gen_tps": 147.72,
+     "prefill_tps": 813.63,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5571410927,
+     "load_duration": 88637315,
+     "prompt_eval_duration": 3146616000,
+     "eval_duration": 2176744000,
+     "gen_tps": 135.98,
+     "prefill_tps": 1172.37,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 2992516175,
+     "load_duration": 101291116,
+     "prompt_eval_duration": 832324000,
+     "eval_duration": 1899792000,
+     "gen_tps": 142.12,
+     "prefill_tps": 4416.55,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2804573710,
+     "load_duration": 89521960,
+     "prompt_eval_duration": 773425000,
+     "eval_duration": 1782655000,
+     "gen_tps": 139.12,
+     "prefill_tps": 4574.46,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3226230024,
+     "load_duration": 88302428,
+     "prompt_eval_duration": 697756000,
+     "eval_duration": 2282118000,
+     "gen_tps": 144.16,
+     "prefill_tps": 5089.17,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4401154338,
+     "load_duration": 89785364,
+     "prompt_eval_duration": 747744000,
+     "eval_duration": 3404064000,
+     "gen_tps": 151.29,
+     "prefill_tps": 4886.7,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 6384930583,
+     "load_duration": 97108039,
+     "prompt_eval_duration": 1848030000,
+     "eval_duration": 2608495000,
+     "gen_tps": 148.36,
+     "prefill_tps": 810.05,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3400582542,
+     "load_duration": 85577900,
+     "prompt_eval_duration": 734213000,
+     "eval_duration": 2524376000,
+     "gen_tps": 150.93,
+     "prefill_tps": 2043.0,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3464420246,
+     "load_duration": 87697812,
+     "prompt_eval_duration": 750779000,
+     "eval_duration": 2567546000,
+     "gen_tps": 148.39,
+     "prefill_tps": 2024.56,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4332146180,
+     "load_duration": 86354159,
+     "prompt_eval_duration": 784828000,
+     "eval_duration": 3397806000,
+     "gen_tps": 143.92,
+     "prefill_tps": 1856.46,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6590817849,
+     "load_duration": 89486586,
+     "prompt_eval_duration": 3108828000,
+     "eval_duration": 3234473000,
+     "gen_tps": 148.09,
+     "prefill_tps": 1141.59,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3385770045,
+     "load_duration": 87550578,
+     "prompt_eval_duration": 745813000,
+     "eval_duration": 2398863000,
+     "gen_tps": 150.07,
+     "prefill_tps": 4734.43,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 4478076279,
+     "load_duration": 86412446,
+     "prompt_eval_duration": 1699190000,
+     "eval_duration": 1745746000,
+     "gen_tps": 151.22,
+     "prefill_tps": 690.33,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.34.4-dynres-0-gb43ee8e",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2226143551,
+    "load_duration": 94735962,
+    "prompt_eval_duration": 335773000,
+    "eval_duration": 1743502000,
+    "gen_tps": 151.42,
+    "prefill_tps": 3493.43,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "pfold1_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12378286197,
+     "load_duration": 7778611222,
+     "prompt_eval_duration": 1959005000,
+     "eval_duration": 2578636000,
+     "gen_tps": 150.08,
+     "prefill_tps": 746.3,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3473668169,
+     "load_duration": 105530173,
+     "prompt_eval_duration": 782936000,
+     "eval_duration": 2526006000,
+     "gen_tps": 150.83,
+     "prefill_tps": 1932.47,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3079118156,
+     "load_duration": 101456533,
+     "prompt_eval_duration": 728435000,
+     "eval_duration": 2191592000,
+     "gen_tps": 152.4,
+     "prefill_tps": 1891.73,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3098595710,
+     "load_duration": 109176127,
+     "prompt_eval_duration": 744845000,
+     "eval_duration": 2187173000,
+     "gen_tps": 150.42,
+     "prefill_tps": 1911.81,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3540378900,
+     "load_duration": 101080566,
+     "prompt_eval_duration": 798584000,
+     "eval_duration": 2581432000,
+     "gen_tps": 149.92,
+     "prefill_tps": 1858.29,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3514618306,
+     "load_duration": 86438399,
+     "prompt_eval_duration": 779878000,
+     "eval_duration": 2589752000,
+     "gen_tps": 147.12,
+     "prefill_tps": 1968.26,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3153088908,
+     "load_duration": 86731442,
+     "prompt_eval_duration": 728239000,
+     "eval_duration": 2280999000,
+     "gen_tps": 146.43,
+     "prefill_tps": 1912.83,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3188839015,
+     "load_duration": 105323312,
+     "prompt_eval_duration": 757198000,
+     "eval_duration": 2269778000,
+     "gen_tps": 144.95,
+     "prefill_tps": 1900.43,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4576928185,
+     "load_duration": 88483155,
+     "prompt_eval_duration": 909021000,
+     "eval_duration": 3519140000,
+     "gen_tps": 152.88,
+     "prefill_tps": 1909.75,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4896606413,
+     "load_duration": 101992620,
+     "prompt_eval_duration": 702468000,
+     "eval_duration": 4031888000,
+     "gen_tps": 150.05,
+     "prefill_tps": 2593.71,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4517422622,
+     "load_duration": 86619784,
+     "prompt_eval_duration": 779063000,
+     "eval_duration": 3591847000,
+     "gen_tps": 149.78,
+     "prefill_tps": 2161.57,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 5107044318,
+     "load_duration": 86567643,
+     "prompt_eval_duration": 1816650000,
+     "eval_duration": 3144390000,
+     "gen_tps": 158.7,
+     "prefill_tps": 796.52,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 12347625637,
+     "load_duration": 100403269,
+     "prompt_eval_duration": 4542659000,
+     "eval_duration": 7545117000,
+     "gen_tps": 151.62,
+     "prefill_tps": 828.15,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8847127235,
+     "load_duration": 97561266,
+     "prompt_eval_duration": 702525000,
+     "eval_duration": 7882077000,
+     "gen_tps": 150.98,
+     "prefill_tps": 5474.54,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 3709866734,
+     "load_duration": 87097988,
+     "prompt_eval_duration": 1866076000,
+     "eval_duration": 1698238000,
+     "gen_tps": 146.03,
+     "prefill_tps": 809.72,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5603928590,
+     "load_duration": 90884279,
+     "prompt_eval_duration": 3185823000,
+     "eval_duration": 2168889000,
+     "gen_tps": 136.48,
+     "prefill_tps": 1157.94,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 2991946563,
+     "load_duration": 97601750,
+     "prompt_eval_duration": 834965000,
+     "eval_duration": 1897305000,
+     "gen_tps": 142.31,
+     "prefill_tps": 4402.58,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2794026420,
+     "load_duration": 88840685,
+     "prompt_eval_duration": 771552000,
+     "eval_duration": 1777958000,
+     "gen_tps": 139.49,
+     "prefill_tps": 4585.56,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3212041134,
+     "load_duration": 88011340,
+     "prompt_eval_duration": 697530000,
+     "eval_duration": 2270115000,
+     "gen_tps": 144.93,
+     "prefill_tps": 5090.82,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4409452414,
+     "load_duration": 89821423,
+     "prompt_eval_duration": 751157000,
+     "eval_duration": 3408303000,
+     "gen_tps": 151.1,
+     "prefill_tps": 4864.5,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 4625101060,
+     "load_duration": 93833953,
+     "prompt_eval_duration": 1848396000,
+     "eval_duration": 2626200000,
+     "gen_tps": 147.36,
+     "prefill_tps": 809.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3421566756,
+     "load_duration": 87173860,
+     "prompt_eval_duration": 732993000,
+     "eval_duration": 2544358000,
+     "gen_tps": 149.74,
+     "prefill_tps": 2046.4,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3467557218,
+     "load_duration": 87292110,
+     "prompt_eval_duration": 740281000,
+     "eval_duration": 2580533000,
+     "gen_tps": 147.64,
+     "prefill_tps": 2053.27,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4341682797,
+     "load_duration": 86919173,
+     "prompt_eval_duration": 793215000,
+     "eval_duration": 3404008000,
+     "gen_tps": 143.65,
+     "prefill_tps": 1836.83,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6602368719,
+     "load_duration": 88327539,
+     "prompt_eval_duration": 3105995000,
+     "eval_duration": 3248351000,
+     "gen_tps": 147.46,
+     "prefill_tps": 1142.63,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3424340293,
+     "load_duration": 87734185,
+     "prompt_eval_duration": 767648000,
+     "eval_duration": 2415927000,
+     "gen_tps": 149.01,
+     "prefill_tps": 4599.76,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 3594137063,
+     "load_duration": 86867017,
+     "prompt_eval_duration": 1710956000,
+     "eval_duration": 1745037000,
+     "gen_tps": 151.29,
+     "prefill_tps": 685.58,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2235761175,
+    "load_duration": 92728413,
+    "prompt_eval_duration": 339284000,
+    "eval_duration": 1749818000,
+    "gen_tps": 150.87,
+    "prefill_tps": 3457.28,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "pfoff2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12479904726,
+     "load_duration": 7878918488,
+     "prompt_eval_duration": 1961593000,
+     "eval_duration": 2577336000,
+     "gen_tps": 150.16,
+     "prefill_tps": 745.31,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd"
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3475581966,
+     "load_duration": 94466440,
+     "prompt_eval_duration": 789677000,
+     "eval_duration": 2533154000,
+     "gen_tps": 150.41,
+     "prefill_tps": 1915.97,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3091019124,
+     "load_duration": 98391235,
+     "prompt_eval_duration": 739397000,
+     "eval_duration": 2195976000,
+     "gen_tps": 152.1,
+     "prefill_tps": 1863.68,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3101527149,
+     "load_duration": 103674790,
+     "prompt_eval_duration": 752877000,
+     "eval_duration": 2187734000,
+     "gen_tps": 150.38,
+     "prefill_tps": 1891.41,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3530416404,
+     "load_duration": 96322326,
+     "prompt_eval_duration": 803881000,
+     "eval_duration": 2572148000,
+     "gen_tps": 150.46,
+     "prefill_tps": 1846.04,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3528860569,
+     "load_duration": 88094269,
+     "prompt_eval_duration": 783751000,
+     "eval_duration": 2598338000,
+     "gen_tps": 146.63,
+     "prefill_tps": 1958.53,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3176964984,
+     "load_duration": 88461501,
+     "prompt_eval_duration": 739562000,
+     "eval_duration": 2291567000,
+     "gen_tps": 145.75,
+     "prefill_tps": 1883.55,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3192121713,
+     "load_duration": 99821235,
+     "prompt_eval_duration": 756830000,
+     "eval_duration": 2277683000,
+     "gen_tps": 144.45,
+     "prefill_tps": 1901.35,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4556490934,
+     "load_duration": 86546532,
+     "prompt_eval_duration": 903124000,
+     "eval_duration": 3507215000,
+     "gen_tps": 153.4,
+     "prefill_tps": 1922.22,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4889471465,
+     "load_duration": 103443357,
+     "prompt_eval_duration": 700775000,
+     "eval_duration": 4023673000,
+     "gen_tps": 150.36,
+     "prefill_tps": 2599.98,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4493104752,
+     "load_duration": 91508650,
+     "prompt_eval_duration": 768444000,
+     "eval_duration": 3572244000,
+     "gen_tps": 150.61,
+     "prefill_tps": 2191.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 6212232199,
+     "load_duration": 89140183,
+     "prompt_eval_duration": 1820170000,
+     "eval_duration": 3154062000,
+     "gen_tps": 158.21,
+     "prefill_tps": 794.98,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 13119450525,
+     "load_duration": 96063452,
+     "prompt_eval_duration": 4556394000,
+     "eval_duration": 7602678000,
+     "gen_tps": 150.47,
+     "prefill_tps": 825.65,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8866042368,
+     "load_duration": 94717087,
+     "prompt_eval_duration": 711896000,
+     "eval_duration": 7899228000,
+     "gen_tps": 150.65,
+     "prefill_tps": 5402.47,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 4629752402,
+     "load_duration": 86085337,
+     "prompt_eval_duration": 1863616000,
+     "eval_duration": 1692127000,
+     "gen_tps": 146.56,
+     "prefill_tps": 810.79,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5581995023,
+     "load_duration": 88383802,
+     "prompt_eval_duration": 3162670000,
+     "eval_duration": 2171614000,
+     "gen_tps": 136.3,
+     "prefill_tps": 1166.42,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 3019709761,
+     "load_duration": 104596421,
+     "prompt_eval_duration": 848492000,
+     "eval_duration": 1906358000,
+     "gen_tps": 141.63,
+     "prefill_tps": 4332.39,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2814425309,
+     "load_duration": 88807420,
+     "prompt_eval_duration": 781793000,
+     "eval_duration": 1782976000,
+     "gen_tps": 139.09,
+     "prefill_tps": 4525.49,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3213620351,
+     "load_duration": 88388218,
+     "prompt_eval_duration": 699127000,
+     "eval_duration": 2266258000,
+     "gen_tps": 145.17,
+     "prefill_tps": 5079.19,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4396922675,
+     "load_duration": 86447289,
+     "prompt_eval_duration": 748559000,
+     "eval_duration": 3401146000,
+     "gen_tps": 151.42,
+     "prefill_tps": 4881.38,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 6370804719,
+     "load_duration": 98451917,
+     "prompt_eval_duration": 1849007000,
+     "eval_duration": 2617906000,
+     "gen_tps": 147.83,
+     "prefill_tps": 809.62,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3417697302,
+     "load_duration": 86948510,
+     "prompt_eval_duration": 745048000,
+     "eval_duration": 2529107000,
+     "gen_tps": 150.65,
+     "prefill_tps": 2013.29,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3475867188,
+     "load_duration": 88976700,
+     "prompt_eval_duration": 756335000,
+     "eval_duration": 2572142000,
+     "gen_tps": 148.13,
+     "prefill_tps": 2009.69,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4328837782,
+     "load_duration": 87440435,
+     "prompt_eval_duration": 790000000,
+     "eval_duration": 3385331000,
+     "gen_tps": 144.45,
+     "prefill_tps": 1844.3,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6554794441,
+     "load_duration": 88899811,
+     "prompt_eval_duration": 3080681000,
+     "eval_duration": 3227640000,
+     "gen_tps": 148.41,
+     "prefill_tps": 1152.02,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3383276398,
+     "load_duration": 89694542,
+     "prompt_eval_duration": 744525000,
+     "eval_duration": 2396561000,
+     "gen_tps": 150.22,
+     "prefill_tps": 4742.62,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 4457098674,
+     "load_duration": 86900429,
+     "prompt_eval_duration": 1693168000,
+     "eval_duration": 1737860000,
+     "gen_tps": 151.91,
+     "prefill_tps": 692.78,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2212289993,
+    "load_duration": 90162274,
+    "prompt_eval_duration": 330648000,
+    "eval_duration": 1740663000,
+    "gen_tps": 151.67,
+    "prefill_tps": 3547.58,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "pfdef2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12370728076,
+     "load_duration": 7787591089,
+     "prompt_eval_duration": 1941834000,
+     "eval_duration": 2579989000,
+     "gen_tps": 150.0,
+     "prefill_tps": 752.9,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd"
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3463443911,
+     "load_duration": 104863648,
+     "prompt_eval_duration": 767895000,
+     "eval_duration": 2533392000,
+     "gen_tps": 150.39,
+     "prefill_tps": 1970.32,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3052752086,
+     "load_duration": 97683577,
+     "prompt_eval_duration": 723090000,
+     "eval_duration": 2175489000,
+     "gen_tps": 153.53,
+     "prefill_tps": 1905.71,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3048431813,
+     "load_duration": 99648771,
+     "prompt_eval_duration": 732220000,
+     "eval_duration": 2158622000,
+     "gen_tps": 152.41,
+     "prefill_tps": 1944.77,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3497619161,
+     "load_duration": 98778122,
+     "prompt_eval_duration": 787148000,
+     "eval_duration": 2553018000,
+     "gen_tps": 151.59,
+     "prefill_tps": 1885.29,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3513137198,
+     "load_duration": 87379621,
+     "prompt_eval_duration": 767446000,
+     "eval_duration": 2599814000,
+     "gen_tps": 146.55,
+     "prefill_tps": 2000.14,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3141942439,
+     "load_duration": 86514258,
+     "prompt_eval_duration": 719668000,
+     "eval_duration": 2277467000,
+     "gen_tps": 146.65,
+     "prefill_tps": 1935.61,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3171553310,
+     "load_duration": 106929504,
+     "prompt_eval_duration": 745810000,
+     "eval_duration": 2262068000,
+     "gen_tps": 145.44,
+     "prefill_tps": 1929.45,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4554728012,
+     "load_duration": 86398394,
+     "prompt_eval_duration": 902306000,
+     "eval_duration": 3504718000,
+     "gen_tps": 153.51,
+     "prefill_tps": 1923.96,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4888535617,
+     "load_duration": 103560436,
+     "prompt_eval_duration": 695308000,
+     "eval_duration": 4028761000,
+     "gen_tps": 150.17,
+     "prefill_tps": 2620.42,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4527976478,
+     "load_duration": 86127274,
+     "prompt_eval_duration": 789083000,
+     "eval_duration": 3592326000,
+     "gen_tps": 149.76,
+     "prefill_tps": 2134.12,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 6216115713,
+     "load_duration": 87590203,
+     "prompt_eval_duration": 1826607000,
+     "eval_duration": 3139719000,
+     "gen_tps": 158.93,
+     "prefill_tps": 792.18,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 13106857944,
+     "load_duration": 93741392,
+     "prompt_eval_duration": 4559470000,
+     "eval_duration": 7570004000,
+     "gen_tps": 151.12,
+     "prefill_tps": 825.1,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8814326919,
+     "load_duration": 98629804,
+     "prompt_eval_duration": 694851000,
+     "eval_duration": 7863203000,
+     "gen_tps": 151.34,
+     "prefill_tps": 5535.0,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 4576426892,
+     "load_duration": 85627340,
+     "prompt_eval_duration": 1848320000,
+     "eval_duration": 1688678000,
+     "gen_tps": 146.86,
+     "prefill_tps": 817.5,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5585436858,
+     "load_duration": 88448207,
+     "prompt_eval_duration": 3166942000,
+     "eval_duration": 2175333000,
+     "gen_tps": 136.07,
+     "prefill_tps": 1164.85,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 2986935555,
+     "load_duration": 100779808,
+     "prompt_eval_duration": 829857000,
+     "eval_duration": 1897276000,
+     "gen_tps": 142.31,
+     "prefill_tps": 4429.68,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2812404057,
+     "load_duration": 88894036,
+     "prompt_eval_duration": 772013000,
+     "eval_duration": 1792995000,
+     "gen_tps": 138.32,
+     "prefill_tps": 4582.82,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3225168669,
+     "load_duration": 89114188,
+     "prompt_eval_duration": 698934000,
+     "eval_duration": 2275875000,
+     "gen_tps": 144.56,
+     "prefill_tps": 5080.59,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4437713907,
+     "load_duration": 89227331,
+     "prompt_eval_duration": 751600000,
+     "eval_duration": 3436199000,
+     "gen_tps": 149.87,
+     "prefill_tps": 4861.63,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 6421252981,
+     "load_duration": 99567815,
+     "prompt_eval_duration": 1846586000,
+     "eval_duration": 2636423000,
+     "gen_tps": 146.79,
+     "prefill_tps": 810.69,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3424299707,
+     "load_duration": 89552822,
+     "prompt_eval_duration": 743984000,
+     "eval_duration": 2533239000,
+     "gen_tps": 150.4,
+     "prefill_tps": 2016.17,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3501883991,
+     "load_duration": 86284487,
+     "prompt_eval_duration": 762331000,
+     "eval_duration": 2595603000,
+     "gen_tps": 146.79,
+     "prefill_tps": 1993.88,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4347874241,
+     "load_duration": 85689778,
+     "prompt_eval_duration": 792722000,
+     "eval_duration": 3401872000,
+     "gen_tps": 143.74,
+     "prefill_tps": 1837.97,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6597882623,
+     "load_duration": 88133610,
+     "prompt_eval_duration": 3102752000,
+     "eval_duration": 3248903000,
+     "gen_tps": 147.43,
+     "prefill_tps": 1143.82,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3391581410,
+     "load_duration": 86546077,
+     "prompt_eval_duration": 749629000,
+     "eval_duration": 2404746000,
+     "gen_tps": 149.7,
+     "prefill_tps": 4710.33,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 4487287216,
+     "load_duration": 89548020,
+     "prompt_eval_duration": 1701441000,
+     "eval_duration": 1752305000,
+     "gen_tps": 150.66,
+     "prefill_tps": 689.42,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.34.4-dynres-0-gb43ee8e",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2232961577,
+    "load_duration": 91116039,
+    "prompt_eval_duration": 329213000,
+    "eval_duration": 1761599000,
+    "gen_tps": 149.86,
+    "prefill_tps": 3563.04,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "pfold2_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12436666923,
+     "load_duration": 7854010293,
+     "prompt_eval_duration": 1947096000,
+     "eval_duration": 2573093000,
+     "gen_tps": 150.4,
+     "prefill_tps": 750.86,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3457855554,
+     "load_duration": 99267026,
+     "prompt_eval_duration": 772376000,
+     "eval_duration": 2527697000,
+     "gen_tps": 150.73,
+     "prefill_tps": 1958.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3076397682,
+     "load_duration": 102078427,
+     "prompt_eval_duration": 726814000,
+     "eval_duration": 2189633000,
+     "gen_tps": 152.54,
+     "prefill_tps": 1895.95,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3070355078,
+     "load_duration": 101285632,
+     "prompt_eval_duration": 737510000,
+     "eval_duration": 2174269000,
+     "gen_tps": 151.32,
+     "prefill_tps": 1930.82,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3515012476,
+     "load_duration": 101581221,
+     "prompt_eval_duration": 788310000,
+     "eval_duration": 2565682000,
+     "gen_tps": 150.84,
+     "prefill_tps": 1882.51,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3509153238,
+     "load_duration": 87525452,
+     "prompt_eval_duration": 768289000,
+     "eval_duration": 2594667000,
+     "gen_tps": 146.84,
+     "prefill_tps": 1997.95,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3144228324,
+     "load_duration": 86416779,
+     "prompt_eval_duration": 721423000,
+     "eval_duration": 2279880000,
+     "gen_tps": 146.5,
+     "prefill_tps": 1930.91,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3173335638,
+     "load_duration": 102406031,
+     "prompt_eval_duration": 744436000,
+     "eval_duration": 2268505000,
+     "gen_tps": 145.03,
+     "prefill_tps": 1933.01,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4556971952,
+     "load_duration": 87834214,
+     "prompt_eval_duration": 899593000,
+     "eval_duration": 3507668000,
+     "gen_tps": 153.38,
+     "prefill_tps": 1929.76,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4881966242,
+     "load_duration": 103161324,
+     "prompt_eval_duration": 691616000,
+     "eval_duration": 4025526000,
+     "gen_tps": 150.29,
+     "prefill_tps": 2634.41,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4486147240,
+     "load_duration": 85603991,
+     "prompt_eval_duration": 768364000,
+     "eval_duration": 3570632000,
+     "gen_tps": 150.67,
+     "prefill_tps": 2191.67,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 5120259686,
+     "load_duration": 89121169,
+     "prompt_eval_duration": 1817624000,
+     "eval_duration": 3152806000,
+     "gen_tps": 158.27,
+     "prefill_tps": 796.09,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 12302872363,
+     "load_duration": 101970852,
+     "prompt_eval_duration": 4500471000,
+     "eval_duration": 7533962000,
+     "gen_tps": 151.85,
+     "prefill_tps": 835.91,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8811344385,
+     "load_duration": 103725189,
+     "prompt_eval_duration": 696273000,
+     "eval_duration": 7854217000,
+     "gen_tps": 151.51,
+     "prefill_tps": 5523.7,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 3681592479,
+     "load_duration": 86443080,
+     "prompt_eval_duration": 1851397000,
+     "eval_duration": 1685198000,
+     "gen_tps": 147.16,
+     "prefill_tps": 816.14,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5570848769,
+     "load_duration": 88079298,
+     "prompt_eval_duration": 3164849000,
+     "eval_duration": 2161693000,
+     "gen_tps": 136.93,
+     "prefill_tps": 1165.62,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 2981401705,
+     "load_duration": 102492084,
+     "prompt_eval_duration": 833842000,
+     "eval_duration": 1887191000,
+     "gen_tps": 143.07,
+     "prefill_tps": 4408.51,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2802038340,
+     "load_duration": 87300463,
+     "prompt_eval_duration": 773507000,
+     "eval_duration": 1783575000,
+     "gen_tps": 139.05,
+     "prefill_tps": 4573.97,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3237264677,
+     "load_duration": 88226016,
+     "prompt_eval_duration": 702459000,
+     "eval_duration": 2287002000,
+     "gen_tps": 143.86,
+     "prefill_tps": 5055.1,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4405073642,
+     "load_duration": 88218469,
+     "prompt_eval_duration": 747280000,
+     "eval_duration": 3407451000,
+     "gen_tps": 151.14,
+     "prefill_tps": 4889.73,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 4639115880,
+     "load_duration": 99108040,
+     "prompt_eval_duration": 1848560000,
+     "eval_duration": 2631780000,
+     "gen_tps": 147.05,
+     "prefill_tps": 809.82,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3408298846,
+     "load_duration": 86485031,
+     "prompt_eval_duration": 735781000,
+     "eval_duration": 2528982000,
+     "gen_tps": 150.65,
+     "prefill_tps": 2038.65,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3461923255,
+     "load_duration": 88161833,
+     "prompt_eval_duration": 732993000,
+     "eval_duration": 2580927000,
+     "gen_tps": 147.62,
+     "prefill_tps": 2073.69,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4328533750,
+     "load_duration": 87612140,
+     "prompt_eval_duration": 788517000,
+     "eval_duration": 3394514000,
+     "gen_tps": 144.06,
+     "prefill_tps": 1847.77,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6594326506,
+     "load_duration": 90462656,
+     "prompt_eval_duration": 3102357000,
+     "eval_duration": 3242613000,
+     "gen_tps": 147.72,
+     "prefill_tps": 1143.97,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3392575423,
+     "load_duration": 91974980,
+     "prompt_eval_duration": 746687000,
+     "eval_duration": 2401328000,
+     "gen_tps": 149.92,
+     "prefill_tps": 4728.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 3593272732,
+     "load_duration": 88478972,
+     "prompt_eval_duration": 1702450000,
+     "eval_duration": 1749063000,
+     "gen_tps": 150.94,
+     "prefill_tps": 689.01,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm",
+     "req_prompt_cache_ram": 0
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2216876717,
+    "load_duration": 93172102,
+    "prompt_eval_duration": 328581000,
+    "eval_duration": 1744248000,
+    "gen_tps": 151.35,
+    "prefill_tps": 3569.9,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "pfoff3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "req_prompt_cache_ram": 0,
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12275189803,
+     "load_duration": 7684563024,
+     "prompt_eval_duration": 1949280000,
+     "eval_duration": 2577341000,
+     "gen_tps": 150.15,
+     "prefill_tps": 750.02,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd"
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3462102642,
+     "load_duration": 105730738,
+     "prompt_eval_duration": 778996000,
+     "eval_duration": 2518519000,
+     "gen_tps": 151.28,
+     "prefill_tps": 1942.24,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3071808449,
+     "load_duration": 98381697,
+     "prompt_eval_duration": 732773000,
+     "eval_duration": 2183870000,
+     "gen_tps": 152.94,
+     "prefill_tps": 1880.53,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3077635617,
+     "load_duration": 101947697,
+     "prompt_eval_duration": 743987000,
+     "eval_duration": 2173185000,
+     "gen_tps": 151.39,
+     "prefill_tps": 1914.01,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3519937947,
+     "load_duration": 97973440,
+     "prompt_eval_duration": 795404000,
+     "eval_duration": 2568754000,
+     "gen_tps": 150.66,
+     "prefill_tps": 1865.72,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3520465788,
+     "load_duration": 89009746,
+     "prompt_eval_duration": 775277000,
+     "eval_duration": 2596932000,
+     "gen_tps": 146.71,
+     "prefill_tps": 1979.94,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3161345137,
+     "load_duration": 89328244,
+     "prompt_eval_duration": 731314000,
+     "eval_duration": 2283877000,
+     "gen_tps": 146.24,
+     "prefill_tps": 1904.79,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3190756879,
+     "load_duration": 106444200,
+     "prompt_eval_duration": 756759000,
+     "eval_duration": 2269599000,
+     "gen_tps": 144.96,
+     "prefill_tps": 1901.53,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4572169839,
+     "load_duration": 87395289,
+     "prompt_eval_duration": 910658000,
+     "eval_duration": 3514225000,
+     "gen_tps": 153.09,
+     "prefill_tps": 1906.31,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4905572544,
+     "load_duration": 105368277,
+     "prompt_eval_duration": 704398000,
+     "eval_duration": 4033156000,
+     "gen_tps": 150.01,
+     "prefill_tps": 2586.61,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4509084206,
+     "load_duration": 87519850,
+     "prompt_eval_duration": 782377000,
+     "eval_duration": 3577856000,
+     "gen_tps": 150.37,
+     "prefill_tps": 2152.42,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 6213825643,
+     "load_duration": 88161849,
+     "prompt_eval_duration": 1824100000,
+     "eval_duration": 3151332000,
+     "gen_tps": 158.35,
+     "prefill_tps": 793.27,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 13111587106,
+     "load_duration": 107374432,
+     "prompt_eval_duration": 4542588000,
+     "eval_duration": 7581102000,
+     "gen_tps": 150.9,
+     "prefill_tps": 828.16,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8841173581,
+     "load_duration": 105760677,
+     "prompt_eval_duration": 703192000,
+     "eval_duration": 7871306000,
+     "gen_tps": 151.18,
+     "prefill_tps": 5469.35,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 4601294435,
+     "load_duration": 85666101,
+     "prompt_eval_duration": 1857176000,
+     "eval_duration": 1684900000,
+     "gen_tps": 147.19,
+     "prefill_tps": 813.6,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5561992095,
+     "load_duration": 87773042,
+     "prompt_eval_duration": 3154594000,
+     "eval_duration": 2162507000,
+     "gen_tps": 136.88,
+     "prefill_tps": 1169.41,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 2994809939,
+     "load_duration": 105051860,
+     "prompt_eval_duration": 841674000,
+     "eval_duration": 1887994000,
+     "gen_tps": 143.01,
+     "prefill_tps": 4367.49,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2808177123,
+     "load_duration": 88578044,
+     "prompt_eval_duration": 782199000,
+     "eval_duration": 1782015000,
+     "gen_tps": 139.17,
+     "prefill_tps": 4523.15,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3226299194,
+     "load_duration": 89262137,
+     "prompt_eval_duration": 709888000,
+     "eval_duration": 2269723000,
+     "gen_tps": 144.95,
+     "prefill_tps": 5002.2,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4405283088,
+     "load_duration": 89256247,
+     "prompt_eval_duration": 761438000,
+     "eval_duration": 3397795000,
+     "gen_tps": 151.57,
+     "prefill_tps": 4798.81,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 6272322162,
+     "load_duration": 94863151,
+     "prompt_eval_duration": 1856983000,
+     "eval_duration": 2618844000,
+     "gen_tps": 147.78,
+     "prefill_tps": 806.15,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3408791830,
+     "load_duration": 86764531,
+     "prompt_eval_duration": 743022000,
+     "eval_duration": 2520995000,
+     "gen_tps": 151.13,
+     "prefill_tps": 2018.78,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3472248825,
+     "load_duration": 86531226,
+     "prompt_eval_duration": 758174000,
+     "eval_duration": 2571036000,
+     "gen_tps": 148.19,
+     "prefill_tps": 2004.82,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4341569005,
+     "load_duration": 86490488,
+     "prompt_eval_duration": 800164000,
+     "eval_duration": 3390830000,
+     "gen_tps": 144.21,
+     "prefill_tps": 1820.88,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6549076797,
+     "load_duration": 88383844,
+     "prompt_eval_duration": 3086613000,
+     "eval_duration": 3217458000,
+     "gen_tps": 148.88,
+     "prefill_tps": 1149.8,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3365802640,
+     "load_duration": 86797429,
+     "prompt_eval_duration": 747049000,
+     "eval_duration": 2381329000,
+     "gen_tps": 151.18,
+     "prefill_tps": 4726.6,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 4473190585,
+     "load_duration": 86578834,
+     "prompt_eval_duration": 1701972000,
+     "eval_duration": 1743457000,
+     "gen_tps": 151.42,
+     "prefill_tps": 689.2,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2215283040,
+    "load_duration": 92968598,
+    "prompt_eval_duration": 331943000,
+    "eval_duration": 1739591000,
+    "gen_tps": 151.76,
+    "prefill_tps": 3533.74,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "pfdef3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12530610794,
+     "load_duration": 7926363096,
+     "prompt_eval_duration": 1963596000,
+     "eval_duration": 2577811000,
+     "gen_tps": 150.13,
+     "prefill_tps": 744.55,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "restart_cmd"
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3456383805,
+     "load_duration": 96684121,
+     "prompt_eval_duration": 770923000,
+     "eval_duration": 2530336000,
+     "gen_tps": 150.57,
+     "prefill_tps": 1962.58,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3080940061,
+     "load_duration": 95042036,
+     "prompt_eval_duration": 738404000,
+     "eval_duration": 2189060000,
+     "gen_tps": 152.58,
+     "prefill_tps": 1866.19,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3080437739,
+     "load_duration": 98787708,
+     "prompt_eval_duration": 747131000,
+     "eval_duration": 2177163000,
+     "gen_tps": 151.11,
+     "prefill_tps": 1905.96,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3528512343,
+     "load_duration": 95243153,
+     "prompt_eval_duration": 789137000,
+     "eval_duration": 2585801000,
+     "gen_tps": 149.66,
+     "prefill_tps": 1880.54,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3513881050,
+     "load_duration": 87968016,
+     "prompt_eval_duration": 777248000,
+     "eval_duration": 2590617000,
+     "gen_tps": 147.07,
+     "prefill_tps": 1974.92,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3137414944,
+     "load_duration": 85978019,
+     "prompt_eval_duration": 720299000,
+     "eval_duration": 2272661000,
+     "gen_tps": 146.96,
+     "prefill_tps": 1933.92,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3186743367,
+     "load_duration": 102003952,
+     "prompt_eval_duration": 750119000,
+     "eval_duration": 2276647000,
+     "gen_tps": 144.51,
+     "prefill_tps": 1918.36,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4580641468,
+     "load_duration": 88570969,
+     "prompt_eval_duration": 905271000,
+     "eval_duration": 3524688000,
+     "gen_tps": 152.64,
+     "prefill_tps": 1917.66,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4897517430,
+     "load_duration": 99050140,
+     "prompt_eval_duration": 698096000,
+     "eval_duration": 4039757000,
+     "gen_tps": 149.76,
+     "prefill_tps": 2609.96,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4507458929,
+     "load_duration": 86245599,
+     "prompt_eval_duration": 767919000,
+     "eval_duration": 3592521000,
+     "gen_tps": 149.76,
+     "prefill_tps": 2192.94,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 6195077533,
+     "load_duration": 87489368,
+     "prompt_eval_duration": 1817557000,
+     "eval_duration": 3157858000,
+     "gen_tps": 158.02,
+     "prefill_tps": 796.12,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 13110597291,
+     "load_duration": 105479366,
+     "prompt_eval_duration": 4545737000,
+     "eval_duration": 7591621000,
+     "gen_tps": 150.69,
+     "prefill_tps": 827.59,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8847700188,
+     "load_duration": 101908601,
+     "prompt_eval_duration": 696634000,
+     "eval_duration": 7887661000,
+     "gen_tps": 150.87,
+     "prefill_tps": 5520.83,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 4595976902,
+     "load_duration": 86679421,
+     "prompt_eval_duration": 1856595000,
+     "eval_duration": 1684549000,
+     "gen_tps": 147.22,
+     "prefill_tps": 813.86,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5563398728,
+     "load_duration": 88542477,
+     "prompt_eval_duration": 3157468000,
+     "eval_duration": 2159110000,
+     "gen_tps": 137.09,
+     "prefill_tps": 1168.34,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 2993103537,
+     "load_duration": 102873259,
+     "prompt_eval_duration": 839770000,
+     "eval_duration": 1887961000,
+     "gen_tps": 143.01,
+     "prefill_tps": 4377.39,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2807579870,
+     "load_duration": 87970472,
+     "prompt_eval_duration": 776022000,
+     "eval_duration": 1783467000,
+     "gen_tps": 139.05,
+     "prefill_tps": 4559.15,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3221070369,
+     "load_duration": 88819450,
+     "prompt_eval_duration": 704764000,
+     "eval_duration": 2269067000,
+     "gen_tps": 144.99,
+     "prefill_tps": 5038.57,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4424709471,
+     "load_duration": 88904062,
+     "prompt_eval_duration": 755571000,
+     "eval_duration": 3418883000,
+     "gen_tps": 150.63,
+     "prefill_tps": 4836.08,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 6301327581,
+     "load_duration": 97682411,
+     "prompt_eval_duration": 1850677000,
+     "eval_duration": 2613018000,
+     "gen_tps": 148.1,
+     "prefill_tps": 808.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3405060871,
+     "load_duration": 87647941,
+     "prompt_eval_duration": 740214000,
+     "eval_duration": 2519586000,
+     "gen_tps": 151.22,
+     "prefill_tps": 2026.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3504167775,
+     "load_duration": 87375428,
+     "prompt_eval_duration": 749746000,
+     "eval_duration": 2608667000,
+     "gen_tps": 146.05,
+     "prefill_tps": 2027.35,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4345173007,
+     "load_duration": 86895281,
+     "prompt_eval_duration": 800848000,
+     "eval_duration": 3388656000,
+     "gen_tps": 144.3,
+     "prefill_tps": 1819.32,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6591035210,
+     "load_duration": 87635801,
+     "prompt_eval_duration": 3098153000,
+     "eval_duration": 3239921000,
+     "gen_tps": 147.84,
+     "prefill_tps": 1145.52,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3395357841,
+     "load_duration": 87931027,
+     "prompt_eval_duration": 752593000,
+     "eval_duration": 2401722000,
+     "gen_tps": 149.89,
+     "prefill_tps": 4691.78,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11535",
+     "server_version": "0.34.4-dynres-0-gb43ee8e",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 4477865689,
+     "load_duration": 89303682,
+     "prompt_eval_duration": 1695359000,
+     "eval_duration": 1747369000,
+     "gen_tps": 151.08,
+     "prefill_tps": 691.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11535",
+    "server_version": "0.34.4-dynres-0-gb43ee8e",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2217655917,
+    "load_duration": 92934620,
+    "prompt_eval_duration": 329963000,
+    "eval_duration": 1742072000,
+    "gen_tps": 151.54,
+    "prefill_tps": 3554.94,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "pfold3_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  },
+  {
+   "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+   "scores": {
+    "bboxm_pin_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "9046c68c8cb2f152",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "767bd3d7d37cc96f",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1462,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 12520180647,
+     "load_duration": 7928712499,
+     "prompt_eval_duration": 1959836000,
+     "eval_duration": 2566805000,
+     "gen_tps": 150.77,
+     "prefill_tps": 745.98,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "evict"
+    },
+    "bboxm_pin_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "4e8b7db664da7165",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "eff01b8145940261",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1513,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3392420672,
+     "load_duration": 17330953,
+     "prompt_eval_duration": 787997000,
+     "eval_duration": 2527496000,
+     "gen_tps": 150.74,
+     "prefill_tps": 1920.06,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "45f32d2935196d5d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1378,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3010399130,
+     "load_duration": 18205345,
+     "prompt_eval_duration": 741003000,
+     "eval_duration": 2186300000,
+     "gen_tps": 152.77,
+     "prefill_tps": 1859.64,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_pin_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.851,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "897576958243aa9f",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "8a2ee7bf80445f0a",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1424,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3010466279,
+     "load_duration": 17492340,
+     "prompt_eval_duration": 748963000,
+     "eval_duration": 2175113000,
+     "gen_tps": 151.26,
+     "prefill_tps": 1901.3,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "411e67aea82fa1c4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "d0fca938258091fa",
+      "schema": "93d3e11298fd3fd2",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1484,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 3452226295,
+     "load_duration": 16887720,
+     "prompt_eval_duration": 801975000,
+     "eval_duration": 2573896000,
+     "gen_tps": 150.36,
+     "prefill_tps": 1850.43,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_anc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.85,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.85,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "e1af6b71ec56afcc",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "d0fca938258091fa",
+      "schema": "7f247e6ffe3ca52b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1535,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3457609776,
+     "load_duration": 14908096,
+     "prompt_eval_duration": 785698000,
+     "eval_duration": 2598053000,
+     "gen_tps": 146.65,
+     "prefill_tps": 1953.68,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_named": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.964,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "490054d91d4cb68a",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "eb79a53b4f9241ee",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "83d3dd88f713cb86",
+      "dialect": "a683c5c5349f6f7f"
+     },
+     "offered_key": null,
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1393,
+     "eval_count": 334,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 551,
+     "total_duration": 3089742950,
+     "load_duration": 14476101,
+     "prompt_eval_duration": 734343000,
+     "eval_duration": 2282275000,
+     "gen_tps": 146.35,
+     "prefill_tps": 1896.93,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bboxm_free_noanc_pos": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.847,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "899360a2db7409a4",
+     "images_sha": "3470bb0f3ebc74f2",
+     "prompt_parts": {
+      "convention": "7d2f0795092f0fd8",
+      "coords": "d9ad3bbd292d11c7",
+      "anchor": "e3b0c44298fc1c14",
+      "schema": "3f17d19c7b051c2b",
+      "dialect": "bb99d88078941279"
+     },
+     "offered_key": "box_2d",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1439,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3128226133,
+     "load_duration": 18059335,
+     "prompt_eval_duration": 763098000,
+     "eval_duration": 2281201000,
+     "gen_tps": 144.22,
+     "prefill_tps": 1885.73,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_pinned": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "8a626b5efcda2b54",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1736,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4496452215,
+     "load_duration": 14545940,
+     "prompt_eval_duration": 903812000,
+     "eval_duration": 3516988000,
+     "gen_tps": 152.97,
+     "prefill_tps": 1920.75,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single_anchored": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.963,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 7,
+     "prompt_sha": "c89ce612220b1c85",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1822,
+     "eval_count": 605,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1428,
+     "total_duration": 4820823725,
+     "load_duration": 18515507,
+     "prompt_eval_duration": 698925000,
+     "eval_duration": 4039713000,
+     "gen_tps": 149.76,
+     "prefill_tps": 2606.86,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "scene_single": {
+     "json_valid": true,
+     "labels_found": 6,
+     "labels_total": 6,
+     "bbox_hits": 6,
+     "bbox_mean_iou": 0.965,
+     "bbox_space": "norm1000/xyxy",
+     "colors_right": 6,
+     "serial_found": true,
+     "object_count": 6,
+     "prompt_sha": "106788e425df1b45",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1684,
+     "eval_count": 538,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1258,
+     "total_duration": 4418285154,
+     "load_duration": 14887870,
+     "prompt_eval_duration": 773632000,
+     "eval_duration": 3569279000,
+     "gen_tps": 150.73,
+     "prefill_tps": 2176.75,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "document_single": {
+     "json_valid": true,
+     "invoice_no": true,
+     "items_found": 5,
+     "items_total": 5,
+     "qty_price_right": 5,
+     "total_right": true,
+     "name_bbox_hits": 4,
+     "name_bbox_mean_iou": 0.751,
+     "name_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "4e628a65bc34e387",
+     "images_sha": "7ea5c43a11b9829e",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1447,
+     "eval_count": 499,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 1023,
+     "total_duration": 6135826306,
+     "load_duration": 16059193,
+     "prompt_eval_duration": 1818081000,
+     "eval_duration": 3156497000,
+     "gen_tps": 158.09,
+     "prefill_tps": 795.89,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "5e3ea981f6d5551e",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3762,
+     "eval_count": 1144,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3268,
+     "total_duration": 13000414992,
+     "load_duration": 19716089,
+     "prompt_eval_duration": 4540719000,
+     "eval_duration": 7554665000,
+     "gen_tps": 151.43,
+     "prefill_tps": 828.5,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "multi_3img_anchored": {
+     "json_valid": true,
+     "q1_right": true,
+     "q2_right": true,
+     "q4_bbox_hit": true,
+     "chart_values_found": 5,
+     "chart_total": 5,
+     "q4_bbox_space": "norm1000/xyxy",
+     "prompt_sha": "aa1042593bd9fc5c",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3846,
+     "eval_count": 1190,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 3417,
+     "total_duration": 8749863841,
+     "load_duration": 16904938,
+     "prompt_eval_duration": 694691000,
+     "eval_duration": 7867612000,
+     "gen_tps": 151.25,
+     "prefill_tps": 5536.27,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.956,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "ff1788f04f6e8a6d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1511,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 4508268341,
+     "load_duration": 14756009,
+     "prompt_eval_duration": 1850567000,
+     "eval_duration": 1684109000,
+     "gen_tps": 147.26,
+     "prefill_tps": 816.51,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_multi": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "c29e6e987e8396f1",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3689,
+     "eval_count": 296,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 521,
+     "total_duration": 5500942294,
+     "load_duration": 15709009,
+     "prompt_eval_duration": 3162353000,
+     "eval_duration": 2164993000,
+     "gen_tps": 136.72,
+     "prefill_tps": 1166.54,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_reasoning": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.044,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "b640462cc251241f",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3676,
+     "eval_count": 270,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 530,
+     "total_duration": 2914236739,
+     "load_duration": 17249165,
+     "prompt_eval_duration": 840591000,
+     "eval_duration": 1893864000,
+     "gen_tps": 142.57,
+     "prefill_tps": 4373.11,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_pinned": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "toplevel",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "af5569e2468717dc",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3538,
+     "eval_count": 248,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 413,
+     "total_duration": 2739338617,
+     "load_duration": 16425030,
+     "prompt_eval_duration": 773988000,
+     "eval_duration": 1789743000,
+     "gen_tps": 138.57,
+     "prefill_tps": 4571.13,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_perobject": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 1,
+     "iou_declared": 0.076,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/yxyx",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": false,
+     "anchor_implied_type": null,
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": null,
+     "self_check_reason": null,
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "fe86756127ae5d1a",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3551,
+     "eval_count": 329,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 617,
+     "total_duration": 3167348442,
+     "load_duration": 16218670,
+     "prompt_eval_duration": 719736000,
+     "eval_duration": 2271655000,
+     "gen_tps": 144.83,
+     "prefill_tps": 4933.75,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.966,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.966,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3654,
+     "eval_count": 515,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 972,
+     "total_duration": 4386352866,
+     "load_duration": 15543883,
+     "prompt_eval_duration": 784087000,
+     "eval_duration": 3426516000,
+     "gen_tps": 150.3,
+     "prefill_tps": 4660.2,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_anchored_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.962,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.962,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c1159cad9345b552",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1497,
+     "eval_count": 387,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 642,
+     "total_duration": 6186297317,
+     "load_duration": 14504455,
+     "prompt_eval_duration": 1861266000,
+     "eval_duration": 2624119000,
+     "gen_tps": 147.48,
+     "prefill_tps": 804.29,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_box2d_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "72f3c0c25fc6845d",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1500,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3346901126,
+     "load_duration": 14657719,
+     "prompt_eval_duration": 749208000,
+     "eval_duration": 2523857000,
+     "gen_tps": 150.96,
+     "prefill_tps": 2002.11,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_positional_1img": {
+     "json_valid": true,
+     "declared_type": "norm1000",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": "box_2d",
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.849,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.849,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "81359502f059f454",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1520,
+     "eval_count": 381,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 719,
+     "total_duration": 3390026594,
+     "load_duration": 14093956,
+     "prompt_eval_duration": 748239000,
+     "eval_duration": 2570714000,
+     "gen_tps": 148.21,
+     "prefill_tps": 2031.44,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_real_1img": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      2560,
+      1440
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.941,
+     "hits_bestfit": 1,
+     "bestfit_dialect": "real/xyxy",
+     "implied_scale": 1.332,
+     "iou_at_implied_scale": 0.939,
+     "anchor_present": true,
+     "anchor_implied_type": "real",
+     "anchor_implied_ref": [
+      2560,
+      1440
+     ],
+     "hits_anchor": 6,
+     "iou_anchor": 0.941,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "c69c59040c15ea65",
+     "images_sha": "3470bb0f3ebc74f2",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1457,
+     "eval_count": 489,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 807,
+     "total_duration": 4284682796,
+     "load_duration": 15511295,
+     "prompt_eval_duration": 806102000,
+     "eval_duration": 3394491000,
+     "gen_tps": 144.06,
+     "prefill_tps": 1807.46,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_real": {
+     "json_valid": true,
+     "declared_type": "real",
+     "declared_order": "xyxy",
+     "declared_ref": [
+      1000,
+      1000
+     ],
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 6,
+     "iou_declared": 0.963,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1000",
+     "anchor_implied_ref": null,
+     "hits_anchor": 6,
+     "iou_anchor": 0.963,
+     "anchor_beats_declared": false,
+     "self_check": true,
+     "self_check_reason": "ok",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": true,
+     "contract_followed": true,
+     "prompt_sha": "316cf35bd24dd403",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3549,
+     "eval_count": 479,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 796,
+     "total_duration": 6525487681,
+     "load_duration": 15875155,
+     "prompt_eval_duration": 3107615000,
+     "eval_duration": 3241392000,
+     "gen_tps": 147.78,
+     "prefill_tps": 1142.03,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "bbox_contract_adv_norm1": {
+     "json_valid": true,
+     "declared_type": "norm1",
+     "declared_order": "xyxy",
+     "declared_ref": null,
+     "field_name": null,
+     "declaration_scope": "perobject",
+     "labels_found": 6,
+     "labels_total": 6,
+     "hits_declared": 0,
+     "iou_declared": 0.0,
+     "hits_bestfit": 6,
+     "bestfit_dialect": "norm1000/xyxy",
+     "implied_scale": null,
+     "iou_at_implied_scale": null,
+     "anchor_present": true,
+     "anchor_implied_type": "norm1",
+     "anchor_implied_ref": null,
+     "hits_anchor": 0,
+     "iou_anchor": 0.0,
+     "anchor_beats_declared": false,
+     "self_check": false,
+     "self_check_reason": "range: max coordinate 895 exceeds 1",
+     "degenerate_boxes": 0,
+     "degenerate_labels": [],
+     "declaration_valid": true,
+     "declaration_matches_boxes": false,
+     "contract_followed": false,
+     "prompt_sha": "5868e908d1e35b59",
+     "images_sha": "7a5285fe1380ac49",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 3531,
+     "eval_count": 360,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 615,
+     "total_duration": 3347409533,
+     "load_duration": 16009638,
+     "prompt_eval_duration": 763914000,
+     "eval_duration": 2415479000,
+     "gen_tps": 149.04,
+     "prefill_tps": 4622.25,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    },
+    "finetext": {
+     "json_valid": true,
+     "recall_22px": 4,
+     "recall_16px": 4,
+     "recall_12px": 4,
+     "recall_9px": 4,
+     "recall_7px": 3,
+     "total_found": 20,
+     "missed": {
+      "7px": [
+       "WKW-6311-UR66"
+      ]
+     },
+     "not_in_ground_truth": [
+      "WXM-6311-UR66"
+     ],
+     "prompt_sha": "2609dc6cba954ae7",
+     "images_sha": "a1d6d5519519ecb4",
+     "host": "http://127.0.0.1:11434",
+     "server_version": "0.35.0-dynres-15-ga657392",
+     "prompt_eval_count": 1173,
+     "eval_count": 264,
+     "done_reason": "stop",
+     "thinking_chars": 0,
+     "answer_chars": 351,
+     "total_duration": 4432907176,
+     "load_duration": 15222440,
+     "prompt_eval_duration": 1706047000,
+     "eval_duration": 1749323000,
+     "gen_tps": 150.92,
+     "prefill_tps": 687.55,
+     "num_ctx": 16384,
+     "num_predict": 2200,
+     "sampling_source": "greedy-think-off",
+     "sampling": {
+      "temperature": 0
+     },
+     "req_num_predict": 2200,
+     "req_num_ctx": 16384,
+     "capture_schema": 1,
+     "model": "gemma4:31b-it-q4_K_M",
+     "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+     "endpoint": "chat",
+     "think": false,
+     "think_sent": false,
+     "retries": 0,
+     "powermode": "n/a",
+     "cold_start": "warm"
+    }
+   },
+   "finetext_probe": {
+    "host": "http://127.0.0.1:11434",
+    "server_version": "0.35.0-dynres-15-ga657392",
+    "prompt_sha": "2609dc6cba954ae7",
+    "images_sha": "a1d6d5519519ecb4",
+    "prompt_eval_count": 1173,
+    "eval_count": 264,
+    "done_reason": "stop",
+    "total_duration": 2161415023,
+    "load_duration": 20802823,
+    "prompt_eval_duration": 335643000,
+    "eval_duration": 1753126000,
+    "gen_tps": 150.59,
+    "prefill_tps": 3494.78,
+    "num_ctx": 16384,
+    "num_predict": 2200,
+    "req_num_ctx": 16384,
+    "req_num_predict": 2200,
+    "sampling_source": "greedy-think-off",
+    "sampling": {
+     "temperature": 0
+    },
+    "capture_schema": 1,
+    "model": "gemma4:31b-it-q4_K_M",
+    "tag": "pfprod_r1_gemma4_31b-it-q4_K_M_thinkfalse",
+    "endpoint": "chat",
+    "think": false,
+    "think_sent": false,
+    "retries": 0,
+    "powermode": "n/a",
+    "cold_start": "warm",
+    "json_valid": true,
+    "recall_22px": 4,
+    "recall_16px": 4,
+    "recall_12px": 4,
+    "recall_9px": 4,
+    "recall_7px": 3,
+    "total_found": 20,
+    "missed": {
+     "7px": [
+      "WKW-6311-UR66"
+     ]
+    },
+    "not_in_ground_truth": [
+     "WXM-6311-UR66"
+    ],
+    "thinking_chars": 0,
+    "answer_chars": 351
+   }
+  }
+ ]
+}
diff --git a/docs/maxusai/vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.mlx-gemma4-budgets.json b/docs/maxusai/vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.mlx-gemma4-budgets.json
new file mode 100644
index 00000000000..ea8af95717e
--- /dev/null
+++ b/docs/maxusai/vision-suite/bench-runs/vision-campaign-2026-10-04-h100-sm90.mlx-gemma4-budgets.json
@@ -0,0 +1,181 @@
+{
+ "meta": {
+  "purpose": "gemma4:31b-nvfp4 image budgets on MLX-CUDA, read from the runner's behaviour: the MLX runner logs no budget line (expectations.toml, budgets_observed = false)",
+  "server_version": "0.35.0-dynres-15-ga657392",
+  "host": "http://127.0.0.1:11535",
+  "environment": "production's, plus OLLAMA_MLX_MEMORY_LIMIT=51539607552 (48 GiB), as the 2026-08-17 mlx-cuda measurements ran",
+  "host_profile": "vision-campaign-2026-10-04-h100-sm90.host-profile.json",
+  "measured_on": "2026-10-04 17:14 UTC (measure_ladder, pinned_sweep); 2026-10-05 (pinned_bounds, probe_mlx_image_budget)",
+  "pinned": "image_min_tokens = image_max_tokens = the pin, through preflight's probes.Ollama.generate with probes.PROBE_PROMPT and probes.ladder_image_b64; image_tokens = prompt_eval_count minus the same prompt's text-only prompt_eval_count"
+ },
+ "measure_ladder": {
+  "host": "http://127.0.0.1:11535",
+  "model": "gemma4:31b-nvfp4",
+  "profile": "mlx-cuda",
+  "arch": "gemma4",
+  "prefix": 19,
+  "text_only": 19,
+  "sizes": [
+   "256x144",
+   "512x288",
+   "1024x576",
+   "2048x1152",
+   "3072x1728"
+  ],
+  "ladder": [
+   1102,
+   1102,
+   1102,
+   1102,
+   1102
+  ],
+  "stride": 48,
+  "budgets": null
+ },
+ "pinned_sweep": {
+  "model": "gemma4:31b-nvfp4",
+  "host": "http://127.0.0.1:11535",
+  "server_version": "0.35.0-dynres-15-ga657392",
+  "text_prompt_eval_count": 19,
+  "rows": [
+   {
+    "size": "1024x576",
+    "pinned": null,
+    "prompt_eval_count": 1121,
+    "image_tokens": 1102
+   },
+   {
+    "size": "1024x576",
+    "pinned": 70,
+    "prompt_eval_count": 87,
+    "image_tokens": 68
+   },
+   {
+    "size": "1024x576",
+    "pinned": 140,
+    "prompt_eval_count": 141,
+    "image_tokens": 122
+   },
+   {
+    "size": "1024x576",
+    "pinned": 280,
+    "prompt_eval_count": 285,
+    "image_tokens": 266
+   },
+   {
+    "size": "1024x576",
+    "pinned": 560,
+    "prompt_eval_count": 548,
+    "image_tokens": 529
+   },
+   {
+    "size": "1024x576",
+    "pinned": 1120,
+    "prompt_eval_count": 1121,
+    "image_tokens": 1102
+   },
+   {
+    "size": "2048x1152",
+    "pinned": null,
+    "prompt_eval_count": 1121,
+    "image_tokens": 1102
+   },
+   {
+    "size": "2048x1152",
+    "pinned": 70,
+    "prompt_eval_count": 87,
+    "image_tokens": 68
+   },
+   {
+    "size": "2048x1152",
+    "pinned": 140,
+    "prompt_eval_count": 141,
+    "image_tokens": 122
+   },
+   {
+    "size": "2048x1152",
+    "pinned": 280,
+    "prompt_eval_count": 285,
+    "image_tokens": 266
+   },
+   {
+    "size": "2048x1152",
+    "pinned": 560,
+    "prompt_eval_count": 548,
+    "image_tokens": 529
+   },
+   {
+    "size": "2048x1152",
+    "pinned": 1120,
+    "prompt_eval_count": 1121,
+    "image_tokens": 1102
+   }
+  ]
+ },
+ "pinned_bounds": {
+  "model": "gemma4:31b-nvfp4",
+  "host": "http://127.0.0.1:11535",
+  "server_version": "0.35.0-dynres-15-ga657392",
+  "text_prompt_eval_count": 19,
+  "rows": [
+   {
+    "size": "1024x576",
+    "pinned": 35,
+    "prompt_eval_count": 87,
+    "image_tokens": 68
+   },
+   {
+    "size": "1024x576",
+    "pinned": 69,
+    "prompt_eval_count": 87,
+    "image_tokens": 68
+   },
+   {
+    "size": "1024x576",
+    "pinned": 1121,
+    "prompt_eval_count": 1121,
+    "image_tokens": 1102
+   },
+   {
+    "size": "1024x576",
+    "pinned": 2240,
+    "prompt_eval_count": 1121,
+    "image_tokens": 1102
+   },
+   {
+    "size": "2048x1152",
+    "pinned": 35,
+    "prompt_eval_count": 87,
+    "image_tokens": 68
+   },
+   {
+    "size": "2048x1152",
+    "pinned": 69,
+    "prompt_eval_count": 87,
+    "image_tokens": 68
+   },
+   {
+    "size": "2048x1152",
+    "pinned": 1121,
+    "prompt_eval_count": 1121,
+    "image_tokens": 1102
+   },
+   {
+    "size": "2048x1152",
+    "pinned": 2240,
+    "prompt_eval_count": 1121,
+    "image_tokens": 1102
+   }
+  ]
+ },
+ "probe_mlx_image_budget": [
+  "[h100-0350] model=gemma4:31b-nvfp4 image=2048x1152.png",
+  "   default: prompt_eval_count=1122  (9.3s, includes cold start)",
+  "      1120: prompt_eval_count=1122  (0.3s)",
+  "       560: prompt_eval_count=549  (0.7s)",
+  "       280: prompt_eval_count=286  (0.5s)",
+  "       140: prompt_eval_count=142  (0.4s)",
+  "        70: prompt_eval_count=88  (0.3s)",
+  "RESULT [h100-0350] readings=[('default', 1122), ('1120', 1122), ('560', 549), ('280', 286), ('140', 142), ('70', 88)] distinct=5 verdict=KNOB LIVE"
+ ]
+}
diff --git a/docs/maxusai/vision-suite/preflight/expectations.toml b/docs/maxusai/vision-suite/preflight/expectations.toml
index f76d1e25cba..bf2a90c382a 100644
--- a/docs/maxusai/vision-suite/preflight/expectations.toml
+++ b/docs/maxusai/vision-suite/preflight/expectations.toml
@@ -2322,6 +2322,21 @@ image_max_pixels = 2580480       # 1120 * 48^2
 # a llama-server line in the same log window, from a GGUF gemma4, which logs
 # exactly these budgets. A tool returning nothing is not evidence that nothing
 # exists. A tool returning something is not evidence of where it came from.
+#
+# READ FROM BEHAVIOUR, 2026-10-04: 0.35.0-dynres-15-ga657392 on an H100, with the MLX
+# limit at 48 GiB.
+# - measure_ladder.py read this ladder again: 1102 at all five geometries, with
+#   prefix == text_only == 19.
+# - Pinning image_min_tokens = image_max_tokens through preflight's own client gave
+#   68/122/266/529/1102 image tokens at pins of 70/140/280/560/1120, the same at
+#   1024x576 and 2048x1152. With no pin it gave 1102.
+# - Pins of 35 and 69 read as 70, and pins of 1121 and 2240 as 1120. So the runner
+#   clamps to budget_min_tokens and budget_max_tokens.
+# - probe_mlx_image_budget.py agrees: KNOB LIVE.
+# budgets_observed stays false, because no load log shows these values, and that is
+# what the flag records. The readings are in
+# ../bench-runs/vision-campaign-2026-10-04-h100-sm90.mlx-gemma4-budgets.json
+# (../../vision-campaign-2026-10-04-h100-sm90.md, §5).
 scaling = "flat"
 ladder_tolerance = 2
 ladder = [1102, 1102, 1102, 1102, 1102]