Medical Reasoning on MedMCQA OOD
66.2AccuracyPRA (Ours)
Evaluation Results
| Method | Links | |
|---|---|---|
| PRA (Ours)Policy=Qwen3-4B-Instruct2026.04 | 66.2 | |
| RAG + SCPolicy=Qwen3-4B-Instruct2026.04 | 64.8 | |
| RAGPolicy=Qwen3-4B-Instruct2026.04 | 63.3 | |
| CoT + SCPolicy=Qwen3-4B-Instruct2026.04 | 62.7 | |
| CoTPolicy=Qwen3-4B-Instruct2026.04 | 61.1 | |
| Direct + SCPolicy=Qwen3-4B-Instruct2026.04 | 55.8 | |
| DirectPolicy=Qwen3-4B-Instruct2026.04 | 55.6 |