List your company

Medical Reasoning on Medbullets OOD

63.6AccuracyPRA (Ours)
48.10452.12756.1560.173Apr 10, 2026

Evaluation results

MethodLinks
PRA (Ours)Policy=Qwen3-4B-Instruct2026.0463.6
CoT + SCPolicy=Qwen3-4B-Instruct2026.0458.7
RAG + SCPolicy=Qwen3-4B-Instruct2026.0458.4
CoTPolicy=Qwen3-4B-Instruct2026.0456.5
RAGPolicy=Qwen3-4B-Instruct2026.0455.7
DirectPolicy=Qwen3-4B-Instruct2026.0448.8
Direct + SCPolicy=Qwen3-4B-Instruct2026.0448.7