Mathematical Reasoning on MATH (MATH Level Scores)
87.3MATH1 ScoreFixed
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| FixedBackbone=Gemma2-9B2026.03 | 87.3 | 74.3 | 62.6 | 38.7 | 22.8 | |
| NoneBackbone=Qwen2.5-7B2026.03 | 87.3 | 80.7 | 71.8 | 57 | 37.6 | |
| MSSRfullBackbone=Gemma2-9B2026.03 | 86.7 | 75.2 | 65.1 | 41.2 | 29.6 | |
| AccuBackbone=Gemma2-9B2026.03 | 86.4 | 73.1 | 65 | 39.2 | 27.6 | |
| MSSRfullBackbone=Qwen2.5-7B2026.03 | 86.4 | 81.7 | 81 | 63.7 | 39.8 | |
| MSSRsplBackbone=Qwen2.5-7B2026.03 | 85.5 | 78.6 | 78.4 | 62.3 | 37.5 | |
| MSSRsplBackbone=Gemma2-9B2026.03 | 85.2 | 74.1 | 63.5 | 39.9 | 26.3 | |
| AccuBackbone=Qwen2.5-7B2026.03 | 84.9 | 82.2 | 75.1 | 59.4 | 37.2 | |
| LossBackbone=Gemma2-9B2026.03 | 84.7 | 72.6 | 63.2 | 39.4 | 25.7 | |
| LossBackbone=Qwen2.5-7B2026.03 | 84.2 | 77.6 | 72.5 | 61.1 | 38 | |
| FixedBackbone=Qwen2.5-7B2026.03 | 84.1 | 78 | 70.1 | 60.6 | 38.3 | |
| MSSRschBackbone=Gemma2-9B2026.03 | 83.4 | 72.5 | 64.9 | 40.9 | 30.2 | |
| MSSRschBackbone=Qwen2.5-7B2026.03 | 82.5 | 79.8 | 76.3 | 59.6 | 38.6 | |
| LossBackbone=Llama-3.1-8B2026.03 | 82.5 | 71.2 | 65.8 | 52.1 | 33.7 | |
| MSSRfullBackbone=Llama-3.1-8B2026.03 | 82.5 | 74.6 | 70.5 | 52.4 | 36.3 | |
| MSSRsplBackbone=Llama-3.1-8B2026.03 | 81.4 | 73.1 | 68.2 | 53 | 33.8 | |
| AccuBackbone=Llama-3.1-8B2026.03 | 81.2 | 72.1 | 67.5 | 51.7 | 35.2 | |
| NoneBackbone=Llama-3.1-8B2026.03 | 80.1 | 70.6 | 61.1 | 47.4 | 30.2 | |
| FixedBackbone=Llama-3.1-8B2026.03 | 78.3 | 71.5 | 67.1 | 51.8 | 32.9 | |
| MSSRschBackbone=Llama-3.1-8B2026.03 | 78.1 | 72.2 | 69.4 | 50.7 | 34.5 | |
| NoneBackbone=Gemma2-9B2026.03 | 77.8 | 68.8 | 64.9 | 42.5 | 24.2 |