Scientific Problem Solving on PhysicsQA
53.22AccuracyReason, Reward, Refine
Evaluation Results
| Method | Links | |
|---|---|---|
| Reason, Reward, RefineBackbone=Phi 3.5 Mini 3.8B2026.07 | 53.22 | |
| Reason, Reward, RefineBackbone=Qwen 2.5 1.5B2026.07 | 49.62 | |
| Reason, Reward, RefineBackbone=LLaMA 3.2 3B2026.07 | 46.73 | |
| RAGBackbone=Phi 3.5 Mini 3.8B2026.07 | 41.59 | |
| DPOBackbone=Phi 3.5 Mini 3.8B2026.07 | 41.49 | |
| FTBackbone=Phi 3.5 Mini 3.8B2026.07 | 39.19 | |
| Reason, Reward, RefineBackbone=LLaMA 3.2 1B2026.07 | 39.02 | |
| RAGBackbone=Qwen 2.5 1.5B2026.07 | 36.91 | |
| CoTBackbone=Phi 3.5 Mini 3.8B2026.07 | 33.35 | |
| RAGBackbone=LLaMA 3.2 3B2026.07 | 31.21 | |
| CoTBackbone=Qwen 2.5 1.5B2026.07 | 30.64 | |
| RAGBackbone=LLaMA 3.2 1B2026.07 | 28.64 | |
| FTBackbone=LLaMA 3.2 3B2026.07 | 27.84 | |
| CoTBackbone=LLaMA 3.2 3B2026.07 | 27.67 | |
| DPOBackbone=LLaMA 3.2 3B2026.07 | 25.95 | |
| DPOBackbone=Qwen 2.5 1.5B2026.07 | 24.32 | |
| FTBackbone=Qwen 2.5 1.5B2026.07 | 23.51 | |
| FTBackbone=LLaMA 3.2 1B2026.07 | 23.24 | |
| CoTBackbone=LLaMA 3.2 1B2026.07 | 21.35 | |
| DPOBackbone=LLaMA 3.2 1B2026.07 | 18.92 |