Clinical Claim Verification on MIMIC-III
79.13AccuracyFact-Checking Module
Evaluation Results
| Method | Links | |
|---|---|---|
| Fact-Checking ModuleBackbone=LLaMA-3.1-8B, Fine-tuning=LoRA2025.12 | 79.13 | |
| Claude-2Extension=DOSSIER2025.12 | 78.62 | |
| Claude-22025.12 | 70.65 | |
| Claude-1Extension=DOSSIER2025.12 | 70.53 | |
| Claude-12025.12 | 66.74 | |
| CodeLlama-13BExtension=DOSSIER2025.12 | 65.76 | |
| CodeLlama-13B2025.12 | 64.13 | |
| T5-EHRSQL2025.12 | 54.56 | |
| MedAlpaca 7B2025.12 | 46.74 | |
| Asclepius 13B2025.12 | 39.26 | |
| Llama2 7B 32kcontext window=32k2025.12 | 32.81 | |
| ClinicalCamel 13B2025.12 | 32.51 |