Risk evaluation in caregiving responses on ADRD-Caregiver Dataset
0.11RubRIXMedalpaca
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| MedalpacaRefinement Stage=Initial2026.01 | 0.11 | — | — | — | |
| MedichatRefinement Stage=Initial2026.01 | 0.08 | — | — | — | |
| QwenRefinement Stage=Initial2026.01 | 0.06 | — | — | — | |
| ClaudeRefinement Stage=Initial2026.01 | 0.05 | — | — | — | |
| Phi-3Refinement Stage=Initial2026.01 | 0.05 | — | — | — | |
| MedalpacaRefinement Stage=Turn 12026.01 | 0.05 | -54.55 | 37.29 | -0.6 | |
| MedalpacaRefinement Stage=Turn 22026.01 | 0.05 | 0 | 2.96 | -0.05 | |
| GPT-4o-miniRefinement Stage=Initial2026.01 | 0.04 | — | — | — | |
| MedichatRefinement Stage=Turn 12026.01 | 0.04 | -50 | 21.16 | -0.38 | |
| QwenRefinement Stage=Turn 12026.01 | 0.03 | -50 | 12.86 | -0.43 | |
| MedichatRefinement Stage=Turn 22026.01 | 0.03 | -25 | 6.92 | -0.12 | |
| QwenRefinement Stage=Turn 22026.01 | 0.02 | -33.33 | 3.7 | -0.12 | |
| Phi-3Refinement Stage=Turn 12026.01 | 0.01 | -80 | 45.37 | -1.14 | |
| Phi-3Refinement Stage=Turn 22026.01 | 0.01 | -18.28 | 2.63 | -0.07 | |
| ClaudeRefinement Stage=Turn 22026.01 | 0.0017 | 6.25 | -0.17 | 0.01 | |
| ClaudeRefinement Stage=Turn 12026.01 | 0.0016 | -96 | 28.75 | -1.53 | |
| GPT-4o-miniRefinement Stage=Turn 12026.01 | 0.0011 | -97.5 | 34.6 | -2.14 | |
| GPT-4o-miniRefinement Stage=Turn 22026.01 | 0.0011 | 0 | 0 | 0 |