Information Extraction on EMR (test)
82.55Jaccard (Case type)Qwen2.5-7b-lora
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Qwen2.5-7b-loraModel Scale=Small Model, Tuning=LoRA2025.12 | 82.55 | 63.21 | 60.35 | |
| Deepseek v3Model Scale=Full-sized Large Model2025.12 | 78.16 | 17.45 | 8.25 | |
| Gemini2.5-proModel Scale=Full-sized Large Model2025.12 | 77.53 | 33.23 | 26.87 | |
| Qwen-maxModel Scale=Full-sized Large Model2025.12 | 74.22 | 25.6 | 15.64 | |
| Claude-4Model Scale=Full-sized Large Model2025.12 | 73.54 | 27.34 | 22.23 | |
| Gpt-5Model Scale=Full-sized Large Model2025.12 | 69.43 | 21.23 | 16.34 |