QA (MAE) on ChatDoctor
0.0396MAETemp Scaling
Evaluation Results
| Method | Links | |
|---|---|---|
| Temp ScalingTarget Ratio (rt)=0.52026.06 | 0.0396 | |
| GRPOTarget Ratio (rt)=0.52026.06 | 0.0535 | |
| DPOTarget Ratio (rt)=0.52026.06 | 0.0567 | |
| SubsamplingTarget Ratio (rt)=0.52026.06 | 0.1271 | |
| No DefenseTarget Ratio (rt)=0.52026.06 | 0.1684 |