Long-form QA on PubMedQA (test)
37.49ROUGE-1Fine-Tuned GPT-4o + MedBioRAG
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Fine-Tuned GPT-4o + MedBioRAGFine-tuned=true, MedBioRAG=true2025.12 | 37.49 | 14.78 | 27.89 | 6.11 | 37.02 | -3.89 | |
| Fine-Tuned GPT-4oFine-tuned=true, MedBioRAG=false2025.12 | 35.82 | 13.55 | 26.09 | 4.34 | 35.33 | -9.23 | |
| GPT-4o + MedBioRAGFine-tuned=false, MedBioRAG=true2025.12 | 26.39 | 9.55 | 17.47 | 2.73 | 18.1 | -7.86 | |
| GPT-4oFine-tuned=false, MedBioRAG=false2025.12 | 25.72 | 9.02 | 17.05 | 2.48 | 17.04 | -9.04 |