Pairwise Comparison Accuracy Refinement on MT-Bench
77.82Original AccuracyLogistic Regression
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Logistic RegressionModel=GPT-5.4-mini, Data=25652026.06 | 77.82 | 74.27 | -3.55 | 77 | |
| Multilayer PerceptronModel=GPT-5.4-mini, Data=25652026.06 | 77.82 | 74.12 | -3.7 | 77 | |
| Random ForestModel=GPT-5.4-mini, Data=25652026.06 | 77.82 | 77.46 | -0.36 | 77 | |
| Label PropagationModel=GPT-5.4-mini, Data=25652026.06 | 77.82 | 64.91 | -12.91 | 77 | |
| nnPUModel=GPT-5.4-mini, Data=25652026.06 | 77.82 | 77.81 | -0.01 | 77 | |
| AURAModel=GPT-5.4-mini, Data=25752026.06 | 77.82 | 81.01 | 3.19 | 87.4 |