Question Answering on GPQA (avg.@8)
43.43Average Score (@8)Qwen2.5-7B-Instruct
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-7B-InstructTraining Pipeline=PSFT → GRPO2025.08 | 43.43 | |
| Qwen2.5-7B-InstructTraining Pipeline=SFT → GRPO2025.08 | 39.39 | |
| Llama3.1-8B-InstructTraining Pipeline=PSFT → GRPO2025.08 | 36.23 | |
| Qwen2.5-7B-InstructTraining Pipeline=PSFT2025.08 | 33.21 | |
| Qwen2.5-7B-InstructTraining Pipeline=SFT2025.08 | 32.89 | |
| Llama3.1-8B-InstructTraining Pipeline=SFT → GRPO2025.08 | 31.06 | |
| Llama3.1-8B-InstructTraining Pipeline=PSFT2025.08 | 26.89 | |
| Llama3.1-8B-InstructTraining Pipeline=SFT2025.08 | 19.38 |