Clinical Response Evaluation on 400 Held-out Clinical Prompts (test)
3.18FactualityPrivMedChat
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| PrivMedChatepsilon=72026.03 | 3.18 | 3.1 | 2.62 | 2.99 | 2.41 | 2.86 | |
| DP-SFTepsilon=72026.03 | 2.93 | 2.57 | 2.36 | 2.38 | 2.14 | 2.48 | |
| SFTepsilon=–2026.03 | 2.76 | 2.53 | 2.32 | 2.36 | 2.04 | 2.4 |