Add deterministic RIMO-N, OlymMATH, and AMO-Bench parser evaluations - #12
Open
taivu1998 wants to merge 2 commits into
Open
Add deterministic RIMO-N, OlymMATH, and AMO-Bench parser evaluations#12taivu1998 wants to merge 2 commits into
taivu1998 wants to merge 2 commits into
Conversation
taivu1998
marked this pull request as ready for review
July 11, 2026 05:55
Keep the benchmark tests active without torch and import all four new tasks in the core lean-install gate.
taivu1998
force-pushed
the
tdv/olympiad-final-answer-benchmarks
branch
from
July 18, 2026 23:46
1d7a0b1 to
42ea1dd
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Add four offline Olympiad final-answer tasks:
RIMONwith 335 exact-string problems,OlymMATHEasyandOlymMATHHardwith 100 MathVerify-graded problems each, and the 39 programmatically gradedAMOBenchParserproblems. Together they add 574 deterministic examples without an LLM judge or proof grader. The shared task base handles boxed-answer extraction, native pass@k aggregation, subgroup metrics, and per-sample grading metadata. Checked-in JSONL files, licenses, source revisions, grader revisions, and SHA-256 hashes keep runtime evaluation network-free and auditable.The task adapters preserve the released protocols: strict extracted-answer matching for RIMO-N, the released English prompt and MathVerify behavior for OlymMATH, and process-isolated SymPy solving with a hard timeout for AMO-Bench variable answers. The tasks participate in Evalchemy's torch-free lean install and per-benchmark extras contract. Backend classification uses the model MRO so task discovery does not import optional OpenAI or vLLM provider classes.
IMOAnswerBench remains deferred while google-deepmind/superhuman#13 and google-deepmind/superhuman#14 require resolution or independent audit. The 11 LLM-judge AMO-Bench problems and CHAMP also remain out of scope until shared judge support and the remaining protocol gates are available.