LLM Ranking on LMArena-based simulation
-2.5Top-8 Ranked CandidatesBradley-Terry ranking
Evaluation Results
| Method | Links | |
|---|---|---|
| Bradley-Terry rankingphi=0.12026.06 | -2.5 | |
| Algorithm 2 rankingphi=0.12026.06 | -2.5 | |
| Algorithm 3 rankingphi=0.12026.06 | -2.5 | |
| Bradley-Terry rankingphi=0.52026.06 | -2.5 | |
| Algorithm 2 rankingphi=0.52026.06 | -2.5 | |
| Algorithm 3 rankingphi=0.52026.06 | -2.5 | |
| Bradley-Terry rankingphi=0.92026.06 | -2.5 | |
| Algorithm 2 rankingphi=0.92026.06 | -2.5 | |
| Algorithm 3 rankingphi=0.92026.06 | -2.5 |