Repository navigation
Benchmark evidence and abstention on 100 source-derived questions - #22
Merged
Merged
Conversation
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Scope
Run 100 deterministic, distinct questions built from the same frozen public CIViC snapshot. Compare context-blind, variant-only, and claim-scoped retrieval/abstention on source-answerability. CI downloads the latest successful refresh artifact, writes an Actions summary and uploads the full EID-level report.
Design
Direct source combinations plus checked counterfactuals: other variant, therapy, cancer, reported effect, and complex profile. Independently enumerate underlying CIViC fields for a source-derived reference; snapshot SHA-256 and actual question mix accompany every run. All truncated trees are excluded uniformly from metric denominators.
Verification
Local Node tests generate and score 100 unique synthetic cases and check baseline error vs scoped abstention. GitHub Actions will run the real 100-question snapshot experiment on this PR.
Limitation
This is an internal source-consistency stress test, not independent expert gold, clinical accuracy or LLM-answer evaluation. No standalone external CIViC MCP baseline is included. The experiment must not be presented as biomedical validation.