Legal Reasoning on JEC-QA Legal Non-MCQ
27.5AccuracyGPT-4o
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-4o2025.10 | 27.5 | |
| CLARITYModel Size=7B2025.10 | 27.2 | |
| Standard RL2025.10 | 26.6 | |
| PRM RL2025.10 | 25.5 | |
| Data Reformulation2025.10 | 25.3 | |
| Qwen2.5Model Size=7B2025.10 | 24.9 | |
| Refine-then-Monitor2025.10 | 23.6 | |
| CLARITYModel Size=3B2025.10 | 22.4 | |
| Qwen2.5Model Size=3B2025.10 | 17.7 |