Add expediency correctness experiment pack#3
Merged
Conversation
b3c3bee to
38f7a5d
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
experiments/<name>/{harness,data,results}with shared repo-level guidanceexperiments/c_vs_d/dataand move its Opus 4.7 public artifacts intoexperiments/c_vs_d/results/opus-4.7src/philosophy_bench/as the shared benchmark package/harness, with bundled C-vs-D data retained as the wheel-install compatibility mirrorResults included
experiments/c_vs_d/results/opus-4.7: original baseline, C-direct, and D-direct judged artifactsexperiments/expediency_vs_correctness/results/opus-4.7: 100/100 judged, axis_mean -0.489, botch_rate 0.06experiments/stability_pairs/results/opus-4.7: 200/200 judged baseline, axis_mean 0.01, botch_rate 0.005experiments/stability_pairs/results/stability_summary.txtandstability_table.jsonsummarize the Opus mirror-pair stability resultsBackward compatibility
experiments/c_vs_d/datafor the original C-vs-D corpussrc/philosophy_bench/dataresults/remain ignored; public artifacts live under each experiment directoryTesting
uv run philosophy-bench scenariosuv run philosophy-bench scenarios --root experiments/c_vs_d/data/scenariosuv run philosophy-bench scenarios --root experiments/expediency_vs_correctness/data/scenariosuv run philosophy-bench scenarios --root experiments/stability_pairs/data/scenariosuv run python experiments/stability_pairs/harness/audit_pairs.py --quietuv run ruff check src tests && uv run ruff format --check src testsuv run --extra dev pytest -quv buildplus wheel install smoke test fortests/test_data_packaging.py