Medical Information Extraction and Classification on MedDialog-CN 1000 labeled samples 1.0 (test)
128.2MCIF ScoreSALP-CG
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SALP-CGModel=gpt-5-mini2025.12 | 128.2 | 98.4 | 99 | 90 | |
| SALP-CGModel=gpt-52025.12 | 116.2 | 99.1 | 98.4 | 90 | |
| SALP-CGModel=qwen3-32b2025.12 | 108.2 | 99.1 | 97.8 | 84 | |
| SALP-CGModel=deepseek-chat2025.12 | 107.3 | 99.7 | 98.9 | 88 | |
| SALP-CGModel=DeepSeek-R1-Distill-Llama-70B2025.12 | 106.3 | 99.7 | 98.7 | 68 | |
| SALP-CGModel=kimi-k2-instruct-09052025.12 | 92.9 | 99.7 | 99 | 83 | |
| SALP-CGModel=llama-3.1-8b-instant2025.12 | 91.7 | 100 | 86.6 | 64 | |
| SALP-CGModel=llama-3.3-70b-versatile2025.12 | 90.9 | 99.5 | 97.3 | 67 | |
| SALP-CGModel=gpt-5-nano2025.12 | 86.9 | 99.2 | 99.3 | 72 | |
| SALP-CGModel=gpt-4o-mini2025.12 | 81.1 | 97.6 | 99.5 | 89 |