Multi-label Classification on NVDRS (test)
89.7Macro F1GPT-5.2
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5.2Approach=Oracle2026.05 | 89.7 | |
| GPT-5.2Approach=Hybrid Prompt2026.05 | 89.3 | |
| GPT-5.2Approach=Complex Prompt2026.05 | 88.3 | |
| Gemini 2.5 ProApproach=Complex Prompt2026.05 | 87.8 | |
| GPT-5.2Approach=Simple Prompt2026.05 | 85.5 | |
| Llama-3 70BApproach=Complex Prompt2026.05 | 83.8 | |
| RoBERTaApproach=Training-based2026.05 | 80 |