Question Answering on BioStarsGPT (test)
0Exact MatchBest Unfine-Tuned (BUT) Baseline
Evaluation Results
| Method | Links | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Best Unfine-Tuned (BUT) BaselineFine-tuning=false, Description=Best performing unfine-tuned model from Table 32025.12 | 0 | 25 | 19 | 43 | 26 | 7 | 24 | 27 | 4 | 27 | 95 | 76 | 57 | 49 | |
| LLaMA-3.2-3B (Fine-tuned)Fine-tuning=true, Base Model=LLaMA-3.2-3B, Method Architecture=BioStarsGPT2025.12 | 0 | 27 | 22 | 39 | 39 | 12 | 22 | 37 | 8 | 24 | 95 | 74 | 59 | 15 | |
| Qwen2.5-7B (Fine-tuned)Fine-tuning=true, Base Model=Qwen2.5-7B, Method Architecture=BioStarsGPT2025.12 | 0 | 30 | 26 | 45 | 44 | 16 | 25 | 39 | 10 | 27 | 97 | 77 | 63 | 18 | |
| Gemma (Fine-tuned)Fine-tuning=true, Base Model=Gemma, Method Architecture=BioStarsGPT2025.12 | 0 | 17 | 19 | 40 | 25 | 8 | 13 | 16 | 4 | 27 | 97 | 74 | 60 | 0 |