Repository navigation
Publish measured Jev follow-up with positive completion recall - #44
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The first frozen Jev run now has all 64 responses, but zero false completion alone hid a serious limitation: images plus telemetry recognized only 1/8 completed physical states. This adds a separate measured follow-up article, every typed result and an offline viewer that makes positive-case recall prominent. The original protocol and unrun release remain unchanged.
The article reports 32 paired native scenes, all 64 single-attempt responses, input-condition-specific reference accuracy, abstention, false completion, positive recognition and measured latency. All four occluded finished scenes fail under both input conditions. Median hosted latency is 1.776 seconds and p95 is 3.287 seconds; this is passive evaluation, not an actuator loop or learned skill.
Evidence includes the frozen native states/images, unchanged evaluation plan, sanitized measured results, an independent run review, native verifier, MIT notices and complete file inventories. No raw provider response bodies or credentials are published. The frozen generator/client/evaluator remains
19acf0734d63ecebc07140c8d84302a54baee4c4; publication source isa3e9afb59f3a9cd7da617cf61416e24fad0a0a5c.Validation: 72 research tests; catalog build/link checks; 11 report and 13 publisher tests; independent reconstruction of all 64 frozen requests and result/metric arithmetic; browser checks for input toggles, all 8 positive cases, recall banner and current/goal comparisons. Existing 26 article bodies and original source evidence are preserved. HF publishes to the new image-goals/v02 namespace and verifies every uploaded byte while retaining previous history.