Mathematical Reasoning on CollegeMath (MV, EM, PK)
52.56Mean Value (MV)I-PPO
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| I-PPOBackbone=Qwen2.5 3B2026.04 | 52.56 | 43.99 | 73.5 | |
| I-PPOBackbone=Phi3 4B2026.04 | 49.06 | 43.82 | 65.44 | |
| I-PPOBackbone=Gemma2 2B2026.04 | 48.38 | 36.44 | 62.62 | |
| I-PPOBackbone=LLaMa3 8B2026.04 | 48.31 | 44.15 | 66.17 | |
| PPOBackbone=Qwen2.5 3B2026.04 | 48.18 | 34.4 | 69.94 | |
| PPOBackbone=Phi3 4B2026.04 | 45.7 | 34.1 | 63.84 | |
| PPOBackbone=Gemma2 2B2026.04 | 44.78 | 28.33 | 60.54 | |
| PPOBackbone=LLaMa3 8B2026.04 | 43.52 | 32.22 | 60.93 | |
| BaselineBackbone=Qwen2.5 3B2026.04 | 37.22 | 26.31 | 63.13 | |
| BaselineBackbone=Phi3 4B2026.04 | 35.59 | 29.22 | 59.01 | |
| BaselineBackbone=Gemma2 2B2026.04 | 34.17 | 20.29 | 55.07 | |
| I-PPOBackbone=Rho 1B2026.04 | 31.25 | 28.17 | 47.89 | |
| BaselineBackbone=LLaMa3 8B2026.04 | 30.77 | 20.59 | 56.81 | |
| PPOBackbone=Rho 1B2026.04 | 28.46 | 21.85 | 45.29 | |
| BaselineBackbone=Rho 1B2026.04 | 17.03 | 10.46 | 29.66 |