Rare disease mention extraction on RareDis n=1,011 docs (test)
84.5Micro-F1Llama-3 70B
Evaluation results
| Method | Links | |
|---|---|---|
| Llama-3 70BApproach=RDMA2025.07 | 84.5 | |
| Qwen3 32BApproach=RDMA2025.07 | 82.8 | |
| GPT-5Approach=RAG2025.07 | 82.6 | |
| Mistral 24BApproach=RDMA2025.07 | 81.4 | |
| GPT-5Approach=Zeroshot2025.07 | 81 | |
| GPT-5Approach=RDMA2025.07 | 78 | |
| Qwen3 32BApproach=RAG2025.07 | 74.8 | |
| BioBERT-MRCApproach=Baselines2025.07 | 71.7 | |
| Llama-3 70BApproach=Zeroshot2025.07 | 70.5 | |
| Mistral 24BApproach=Zeroshot2025.07 | 70.4 | |
| BioClinicalBERTApproach=Baselines2025.07 | 67.3 | |
| Qwen3 32BApproach=Zeroshot2025.07 | 65.1 | |
| Llama-3 8BApproach=RDMA2025.07 | 59.5 | |
| DictionaryApproach=Baselines2025.07 | 55.5 | |
| Llama-3 8BApproach=Zeroshot2025.07 | 54.8 | |
| Llama-3 70BApproach=RAG2025.07 | 48.3 | |
| Llama-3 8BApproach=RAG2025.07 | 26.7 | |
| i2b2Approach=Baselines2025.07 | 26 | |
| Mistral 24BApproach=RAG2025.07 | 22.6 |