feat(extradup): add kill-scan CLI for the frozen ExtraDup corpus - #336
Merged
Conversation
python -m openadapt_evals.extradup kill-scan runs the MockMed proof rewards (visual_only vs certified_sor) and prints silent-incorrect-success on gold-FAIL mutants vs honest-write on the control. execute_seal and production_seal stay false. --verdicts scores another checker on the same cells. Mutants stay off the training reward.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
ExtraDup kill-scan
python -m openadapt_evals.extradup kill-scanis the one command for the frozen ExtraDup corpus.It scores
visual_onlyvscertified_soron the same MockMed rollouts the 2026-09-01 proof already published: 15/15 gold-FAIL paid by the visual reward, 0/15 FAIL and 3/3 honest on the certified reward. Stdout prints silent-incorrect-success on the mutants and honest-write on the control.execute_sealandproduction_sealstay false. Mutants stay off the training reward.--verdicts path.jsonscores a paid/not-paid map from another checker or agent on those cells.--dump-corpuswrites the frozen snapshots.python -m openadapt_evals.reward.proofstill generates the committed table. This wraps it so you don't have to start at docs/reward to find the command.Opened by an agent session, not the founder.