Text Generation on MIMIC-BHC 100 samples (test)
3.973MedHelm Jury ScoreCAPT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CAPTComponents=Mnew + Mold-clin + Mold2026.01 | 3.973 | 31 | |
| Qwen3-30BModel Category=New-gen general (Mnew)2026.01 | 3.932 | 30 | |
| UniTEComponents=Mnew + Mold-clin2026.01 | 3.747 | 20 | |
| Proxy TuningComponents=Mold-L + Mold-clin + Mold2026.01 | 3.456 | 8 | |
| LLaMA-2-70B-chatModel Category=Old-gen large general (Mold-L)2026.01 | 3.323 | 13 | |
| MeLLaMA-13B-chatModel Category=Old-gen clinical (Mold-clin)2026.01 | 2.906 | 14.14 |