Conformal Prediction on Long-form QA landmarks
80.39CoverageOriginal
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| OriginalModel=Gemma-3 12B, Alpha=0.20, Target Coverage=0.802026.04 | 80.39 | 80.67 | |
| Adaptive Conformal PredictionModel=Gemma-3 12B, Alpha=0.20, Target Coverage=0.802026.04 | 79.89 | 80.53 |