Uncertainty Quantification on BBH (test)
90.48AUROCCAGE-CAL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CAGE-CALRollouts=3, Topologies=52026.05 | 90.48 | 91.52 | |
| Answer entropyRollouts=3, Topologies=52026.05 | 67.98 | 84.94 | |
| DAERollouts=3, Topologies=52026.05 | 67.96 | 84.94 | |
| Plurality voteRollouts=3, Topologies=52026.05 | 66.15 | 84.22 | |
| MATURollouts=3, Topologies=52026.05 | 48.81 | 73.89 | |
| Avg-logprobRollouts=3, Topologies=52026.05 | 46.11 | 49.77 |