System Ranking Correlation on CRSArena-Eval Turn-level
0.93Pearson Correlation (r)FACE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| FACE2025.05 | 0.93 | 0.842 | |
| LLMDirect2025.05 | 0.86 | 0.822 | |
| G-Eval2025.05 | 0.74 | 0.84 | |
| Distinct-32025.05 | 0.716 | 0.841 | |
| Distinct-42025.05 | 0.654 | 0.8 | |
| iEvaLMattrUser Simulation=Attribute-based2025.05 | 0.527 | 0.527 | |
| iEvaLMfreeUser Simulation=LLM-based free-form2025.05 | 0.191 | 0.191 | |
| R@102025.05 | -0.192 | 0.048 | |
| R@12025.05 | -0.197 | 0.06 |