System Ranking Correlation on CRSArena-Eval Dial-level
0.913Pearson Correlation (r)FACE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FACE2025.05 | 0.913 | 0.837 | |
| G-Eval2025.05 | 0.893 | 0.83 | |
| LLMDirect2025.05 | 0.872 | 0.799 | |
| Distinct-32025.05 | 0.665 | 0.78 | |
| Distinct-42025.05 | 0.609 | 0.76 | |
| iEvaLMattrUser Simulation=Attribute-based2025.05 | 0.553 | 0.553 | |
| iEvaLMfreeUser Simulation=LLM-based free-form2025.05 | 0.232 | 0.232 | |
| R@102025.05 | -0.111 | 0.071 | |
| R@12025.05 | -0.12 | 0.081 |