Note Generation on MM-HEALTHCN (Overall)
81.17Note Helpfulness (H)EVONOTE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| EVONOTEBackbone=MedGemma-27B2026.06 | 81.17 | 89.59 | |
| EVONOTE-FMBackbone=MedGemma-27B, Memory=frozen warm-up2026.06 | 73.17 | 84.75 | |
| DeepResearch2026.06 | 72.33 | 78.09 | |
| EVONOTE-NMBackbone=MedGemma-27B, Memory=removed retrieval and update2026.06 | 71.83 | 84.25 | |
| ReMemBackbone=MedGemma-27B2026.06 | 69.75 | 74.5 | |
| Grok-1.52026.06 | 69.25 | 76.92 | |
| EVONOTE-NM-NABackbone=MedGemma-27B, Memory=removed retrieval and update, Claim Analyzer=removed2026.06 | 67 | 79.42 | |
| ExpRAGBackbone=MedGemma-27B2026.06 | 65.75 | 79.63 | |
| CrowdNotes+Backbone=MedGemma-27B2026.06 | 64.92 | 69.21 | |
| Gemini-1.5-Flash2026.06 | 50.25 | 74.96 | |
| Human Baseline2026.06 | 38.33 | — | |
| GPT-4o2026.06 | 26.92 | 65.96 |