Bias quantification on Us vs. Them Critical texts
0.536Pearson Correlation Coefficient (r)mini_listwiseBT
Evaluation results
| Method | Links | |
|---|---|---|
| mini_listwiseBTModel=GPT-5-mini, Rating Technique=Listwise Bradley-Terry estimation2025.12 | 0.536 | |
| mini_directModel=GPT-5-mini, Rating Technique=Direct assessment2025.12 | 0.505 | |
| mini_listwiseEloModel=GPT-5-mini, Rating Technique=Listwise Elo rating2025.12 | 0.5 | |
| nano_listwiseBTModel=GPT-5-nano, Rating Technique=Listwise Bradley-Terry estimation2025.12 | 0.46 | |
| nano_listwiseEloModel=GPT-5-nano, Rating Technique=Listwise Elo rating2025.12 | 0.431 |