gpqa_mcq
gpqa_mcq
¶
GPQA MCQ scorer — LLM-based letter extraction + exact match.
Adapted from IPW's mcq.py and gpqa.py evaluation handlers.
Classes¶
GPQAScorer
¶
GPQAScorer(judge_backend: InferenceBackend, judge_model: str)
Bases: LLMJudgeScorer
Score GPQA responses by extracting answer letter via LLM.