Paraphrasing and Simplification of radiological findings on MIMIC
38.3Para ScoreLlama 3.3
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Llama 3.3size=70B, sampling=top-p=0.9, T=12026.06 | 38.3 | 26 | |
| Qwen3size=8B, sampling=top-p=0.9, T=12026.06 | 31.1 | 18.2 | |
| Claude 4.5version=H, sampling=top-p=0.9, T=12026.06 | 29.1 | 20.1 | |
| gpt-5-nanosampling=top-p=0.9, T=12026.06 | 25.8 | 12.5 | |
| Gemini 3.1version=fl, sampling=top-p=0.9, T=12026.06 | 24.3 | 26.8 |