Scientific Reasoning on Theorem QA (accuracy)
66.4Scientific Reasoning QA AccuracyCritique-GRPO
Evaluation results
| Method | Links | |
|---|---|---|
| Critique-GRPOCritique Model=GPT-4o, Decoding budget=16,384 tokens2025.06 | 66.4 | |
| R1-GRPOCritique Model=None, Decoding budget=16,384 tokens2025.06 | 64 | |
| Qwen3-32BCritique Model=None, Decoding budget=16,384 tokens2025.06 | 62.4 |