Bias quantification on Us vs. Them (Full dataset)
0.542Pearson Correlation Coefficientmini_direct
Evaluation Results
| Method | Links | |
|---|---|---|
| mini_directModel=GPT-5-mini, Rating Technique=Direct assessment2025.12 | 0.542 | |
| mini_listwiseBTModel=GPT-5-mini, Rating Technique=Listwise Bradley-Terry estimation2025.12 | 0.537 | |
| mini_listwiseEloModel=GPT-5-mini, Rating Technique=Listwise Elo rating2025.12 | 0.51 | |
| nano_listwiseEloModel=GPT-5-nano, Rating Technique=Listwise Elo rating2025.12 | 0.349 | |
| nano_listwiseBTModel=GPT-5-nano, Rating Technique=Listwise Bradley-Terry estimation2025.12 | 0.346 |