Medical Question Answering on EHRNoteQA
98Calibrated AccuracyQwen3-8B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3-8BEvaluation Protocol=Med-HEAL2026.05 | 98 | 1.7 | 0.07 | |
| Qwen3-8BEvaluation Protocol=Zero-shot2026.05 | 96.3 | — | — | |
| Llama-3.1-8BEvaluation Protocol=Med-HEAL2026.05 | 94.1 | 2 | 0.024 | |
| Qwen2.5-7BEvaluation Protocol=Med-HEAL2026.05 | 92.8 | 1.2 | 0.02 | |
| Llama-3.1-8BEvaluation Protocol=Zero-shot2026.05 | 92.1 | — | — | |
| Qwen2.5-7BEvaluation Protocol=Zero-shot2026.05 | 91.6 | — | — | |
| DeepSeek-R1-8BEvaluation Protocol=Med-HEAL2026.05 | 85.9 | 9 | 0.001 | |
| DeepSeek-R1-8BEvaluation Protocol=Zero-shot2026.05 | 76.9 | — | — | |
| BioMistral-7BEvaluation Protocol=Med-HEAL2026.05 | 50.4 | 2.3 | 0.11 | |
| BioMistral-7BEvaluation Protocol=Zero-shot2026.05 | 48.1 | — | — |