STEM Reasoning on GPQA Diamond (AUROC)
82.1AUROCSCVC
Evaluation Results
| Method | Links | |
|---|---|---|
| SCVCSampling budget (K)=82026.03 | 82.1 | |
| SCVCSampling budget (K)=52026.03 | 81.6 | |
| VCSampling budget (K)=82026.03 | 80.8 | |
| SCVCSampling budget (K)=22026.03 | 80.3 | |
| VCSampling budget (K)=52026.03 | 80.2 | |
| VCSampling budget (K)=22026.03 | 76.8 | |
| SCSampling budget (K)=82026.03 | 75.4 | |
| VCSampling budget (K)=12026.03 | 74.6 | |
| SCSampling budget (K)=52026.03 | 73.3 | |
| SCSampling budget (K)=22026.03 | 66.3 |