Scientific Question Answering on GPQA Diamond (pass@50)
92.42pass@50Gemma-3-27b-it
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemma-3-27b-itSampling Strategy=GUIDEDSAMPLING2025.10 | 92.42 | |
| Gemma-3-27b-itSampling Strategy=Repeated Sampling2025.10 | 70.2 |
Loading the SOTA2 catalog…