Expert Preference Pairwise on LMArena Expert Med
0.476Kendall's tau_bIRT(All) Ranking
Evaluation Results
| Method | Links | |
|---|---|---|
| IRT(All) Rankingn=21, Domain (Format)=Exp Prefer (Pairwise)2025.09 | 0.476 | |
| IRT (pr.) Rankingn=21, Domain (Format)=Exp Prefer (Pairwise)2025.09 | 0.476 | |
| Accuracy Rankingn=21, Domain (Format)=Exp Prefer (Pairwise)2025.09 | 0.438 |