Factuality Evaluation on Medical Empathy Editing Dataset (supplementary evaluation)
934Original Facts CountGemini-2.5-Flash
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Gemini-2.5-Flashediting_strategy=empathy-editing2026.01 | 934 | 855 | 1,194 | 1,081 | 113 | 8.5 | 9.5 | |
| Llama-3.1-70Bparameters=70B, editing_strategy=empathy-editing2026.01 | 934 | 743 | 1,230 | 970 | 260 | 20.4 | 21.1 | |
| Mistral-7Bparameters=7B, editing_strategy=empathy-editing2026.01 | 934 | 732 | 1,373 | 1,053 | 320 | 21.6 | 23.3 | |
| Qwen3-14Bparameters=14B, editing_strategy=empathy-editing2026.01 | 934 | 744 | 1,429 | 1,074 | 355 | 20.3 | 24.8 |