Mathematical Reasoning on SVAMP 8-shot (test)
92AccuracySFT
Evaluation results
| Method | Links | |
|---|---|---|
| SFTBackbone=Qwen2.5-7B-Instruct, Training=SFT, Few-shot=8-shot2026.03 | 92 | |
| MIPOBackbone=Qwen2.5-7B-Instruct, Training=MIPO, Few-shot=8-shot2026.03 | 91.67 | |
| MIPOBackbone=Qwen2.5-3B-Instruct, Training=MIPO, Few-shot=8-shot2026.03 | 91.33 | |
| Qwen2.5-7B-InstructBackbone=Qwen2.5-7B-Instruct, Few-shot=8-shot2026.03 | 91.33 | |
| SFTBackbone=Qwen2.5-7B-Instruct, Training=SFT, Ground-truth usage=true, Few-shot=8-shot2026.03 | 91.33 | |
| RLVRBackbone=Qwen2.5-7B-Instruct, Training=RLVR, Ground-truth usage=true, Few-shot=8-shot2026.03 | 91.33 | |
| Qwen2.5-3B-InstructBackbone=Qwen2.5-3B-Instruct, Few-shot=8-shot2026.03 | 90.67 | |
| SFTBackbone=Qwen2.5-3B-Instruct, Training=SFT, Ground-truth usage=true, Few-shot=8-shot2026.03 | 90.56 | |
| RLVRBackbone=Qwen2.5-3B-Instruct, Training=RLVR, Ground-truth usage=true, Few-shot=8-shot2026.03 | 90.22 | |
| SFTBackbone=Qwen2.5-3B-Instruct, Training=SFT, Few-shot=8-shot2026.03 | 89.78 | |
| RLVRBackbone=Qwen2.5-1.5B-Instruct, Training=RLVR, Ground-truth usage=true, Few-shot=8-shot2026.03 | 83 | |
| Qwen2.5-1.5B-InstructBackbone=Qwen2.5-1.5B-Instruct, Few-shot=8-shot2026.03 | 82.33 | |
| SFTBackbone=Qwen2.5-1.5B-Instruct, Training=SFT, Few-shot=8-shot2026.03 | 82 | |
| MIPOBackbone=Qwen2.5-1.5B-Instruct, Training=MIPO, Few-shot=8-shot2026.03 | 81.67 | |
| SFTBackbone=Qwen2.5-1.5B-Instruct, Training=SFT, Ground-truth usage=true, Few-shot=8-shot2026.03 | 81.33 | |
| SFTBackbone=Llama-3.2-3B-Instruct, Training=SFT, Few-shot=8-shot2026.03 | 79.66 | |
| RLVRBackbone=Llama-3.2-3B-Instruct, Training=RLVR, Ground-truth usage=true, Few-shot=8-shot2026.03 | 79.11 | |
| Llama-3.2-3B-InstructBackbone=Llama-3.2-3B-Instruct, Few-shot=8-shot2026.03 | 78.67 | |
| MIPOBackbone=Llama-3.2-3B-Instruct, Training=MIPO, Few-shot=8-shot2026.03 | 78.22 | |
| SFTBackbone=Llama-3.2-3B-Instruct, Training=SFT, Ground-truth usage=true, Few-shot=8-shot2026.03 | 77.21 | |
| MIPOBackbone=Llama-3.2-1B-Instruct, Training=MIPO, Few-shot=8-shot2026.03 | 60.11 | |
| SFTBackbone=Llama-3.2-1B-Instruct, Training=SFT, Few-shot=8-shot2026.03 | 54.44 | |
| SFTBackbone=Llama-3.2-1B-Instruct, Training=SFT, Ground-truth usage=true, Few-shot=8-shot2026.03 | 52.56 | |
| RLVRBackbone=Llama-3.2-1B-Instruct, Training=RLVR, Ground-truth usage=true, Few-shot=8-shot2026.03 | 52.56 | |
| Llama-3.2-1B-InstructBackbone=Llama-3.2-1B-Instruct, Few-shot=8-shot2026.03 | 51.67 |