Add temperature/repeated-sampling support and Wilson intervals#138
Merged
Conversation
- DeepSeekLLMAnalyst gains a configurable temperature (default 0.0, so existing deterministic runs are unchanged). - run_audit_eval.py accepts --samples-per-task and --temperature, keys the cache and checkpoint by sample (sample 0 keeps the legacy key so the deterministic main-table cache replays), enabling confidence intervals over auditor recall without disturbing the temp-0 results. - statistics.wilson_interval: binomial score interval, robust at the 0/1 extremes that the auditor recalls hit, with unit tests.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Infrastructure for confidence intervals on auditor recall (FinAudit / paper 04):
run_audit_eval.py--samples-per-task/--temperature, cache+checkpoint keyed by sample (sample 0 keeps the legacy key so the temp-0 main table replays)statistics.wilson_intervalwith unit tests (robust at 0/1 extremes)The temp-0.7 3-sample CI experiment is running; results land separately.
Test plan