Mathematical Reasoning on ToxicGSM
89.84AccuracyθSFM (SAFEMATH)
Evaluation Results
| Method | Links | |
|---|---|---|
| θSFM (SAFEMATH)Model Family=LlamaMath, alpha=0.3, beta=0.12026.03 | 89.84 | |
| θSFM (SAFEMATH)Model Family=DeepSeekMath, alpha=0.1, beta=0.12026.03 | 88.23 | |
| θSF (SAFEMATH)Model Family=LlamaMath, beta=0.012026.03 | 88.02 | |
| θSF (SAFEMATH)Model Family=DeepSeekMath, beta=0.22026.03 | 87.86 | |
| θBModel Family=DeepSeekMath2026.03 | 87.76 | |
| θMModel Family=DeepSeekMath2026.03 | 86.98 | |
| θBModel Family=LlamaMath2026.03 | 86.35 | |
| θMModel Family=LlamaMath2026.03 | 84.9 | |
| θSFM (SAFEMATH)Model Family=Qwen2Math, alpha=0.3, beta=0.012026.03 | 80.83 | |
| θSF (SAFEMATH)Model Family=Qwen2Math, beta=0.12026.03 | 78.8 | |
| θMModel Family=Qwen2Math2026.03 | 78.18 | |
| θBModel Family=Qwen2Math2026.03 | 70.89 | |
| θFWModel Family=DeepSeekMath2026.03 | 60.94 | |
| θFWModel Family=LlamaMath2026.03 | 42.14 | |
| θFWModel Family=Qwen2Math2026.03 | 4.06 |