General Domain Reasoning on GPQA (Accuracy)
40.4AccuracyRFT
Evaluation Results
| Method | Links | |
|---|---|---|
| RFTBase Model=Qwen2.5-Math-7B2025.07 | 40.4 | |
| LUFFYBase Model=Qwen2.5-Math-7B2025.07 | 39.9 | |
| Prefix-RFTBase Model=Qwen2.5-Math-7B2025.07 | 39.1 | |
| ReLIFTBase Model=Qwen2.5-Math-7B2025.07 | 37.9 | |
| UFTBase Model=Qwen2.5-Math-7B2025.07 | 34.5 | |
| SFTBase Model=Qwen2.5-Math-7B2025.07 | 24.7 | |
| SFT+RFTBase Model=Qwen2.5-Math-7B2025.07 | 24.2 | |
| Oat-ZeroBase Model=Qwen2.5-Math-7B2025.07 | 23.7 | |
| RL w/ SFT LossBase Model=Qwen2.5-Math-7B2025.07 | 23.7 | |
| SimpleRL-ZeroBase Model=Qwen2.5-Math-7B2025.07 | 23.2 | |
| Qwen2.5-Math-7B2025.07 | 11.1 |