Incidentaloma classification on Incidentaloma classification (test)
99F1 Class 0BioClinicalModernBERT
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| BioClinicalModernBERTModel Type=Supervised Encoder-based Model, Cost-Sensitive Learning (CS)=false2025.12 | 99 | 79 | 61 | 95 | 70 | |
| BioClinicalModernBERTModel Type=Supervised Encoder-based Model, Cost-Sensitive Learning (CS)=true2025.12 | 99 | 72 | 60 | 95 | 66 | |
| ModernBERTModel Type=Supervised Encoder-based Model, Cost-Sensitive Learning (CS)=false2025.12 | 99 | 76 | 60 | 95 | 68 | |
| ModernBERTModel Type=Supervised Encoder-based Model, Cost-Sensitive Learning (CS)=true2025.12 | 99 | 77 | 63 | 95 | 70 | |
| GPT-4oModel Type=Large Language Model, Prompting Strategy=Anatomy-aware2025.12 | 97 | 82 | 71 | 94 | 77 | |
| GPT-OSS-20bModel Type=Large Language Model, Prompting Strategy=Anatomy-aware2025.12 | 97 | 84 | 73 | 94 | 79 | |
| Fine-tuned Llama 3.1-8BModel Type=Large Language Model, Training Protocol=Fine-tuned2025.12 | 96 | 62 | 46 | 91 | 54 | |
| GPT-4oModel Type=Large Language Model, Prompting Strategy=Base2025.12 | 96 | 76 | 62 | 92 | 69 | |
| GPT-OSS-20bModel Type=Large Language Model, Prompting Strategy=Base2025.12 | 96 | 81 | 71 | 93 | 76 | |
| Inter-annotator Agreement (IAA)Type=Human Agreement2025.12 | 93 | 81 | 70 | 89 | 76 | |
| Llama 3.1-8BModel Type=Large Language Model, Prompting Strategy=Anatomy-aware2025.12 | 90 | 61 | 64 | 82 | 63 | |
| Llama 3.1-8BModel Type=Large Language Model, Prompting Strategy=Base2025.12 | 89 | 59 | 46 | 81 | 52 |