Logical Reasoning on ReClor (Pass@1)
61.8Pass@1Qwen2.5-7B + PPO w/ VERL.
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-7B + PPO w/ VERL.Base Model=Qwen2.5-7B, Training=PPO w/ VERL.2025.09 | 61.8 | |
| Qwen2.5-7B + GRPO w/ VERL.Base Model=Qwen2.5-7B, Training=GRPO w/ VERL.2025.09 | 61.6 | |
| Qwen2.5-7B + GRPOBase Model=Qwen2.5-7B, Training=GRPO2025.09 | 58.8 | |
| Qwen2.5-7B + PPOBase Model=Qwen2.5-7B, Training=PPO2025.09 | 58.6 | |
| Qwen2.5-7BBase Model=Qwen2.5-7B2025.09 | 56 | |
| Mathstral-7B-v0.1 + GRPO w/ VERL.Base Model=Mathstral-7B-v0.1, Training=GRPO w/ VERL.2025.09 | 49.7 | |
| Mathstral-7B-v0.1 + PPOBase Model=Mathstral-7B-v0.1, Training=PPO2025.09 | 46.5 | |
| Mathstral-7B-v0.1 + GRPOBase Model=Mathstral-7B-v0.1, Training=GRPO2025.09 | 46.3 | |
| Mathstral-7B-v0.1 + PPO w/ VERL.Base Model=Mathstral-7B-v0.1, Training=PPO w/ VERL.2025.09 | 46.2 | |
| Mathstral-7B-v0.1Base Model=Mathstral-7B-v0.12025.09 | 43.6 |