Knowledge & QA on MMLU-Pro
88.4Accuracy (Knowledge & QA)CoT2-Meta
Evaluation results
| Method | Links | |
|---|---|---|
| CoT2-MetaStrategy=Ours (CoT2-Meta), Inference Budget=C=162026.03 | 88.4 | |
| ReST-MCTS*Strategy=ReST-MCTS*, Inference Budget=C=162026.03 | 86.1 | |
| Vanilla ToTStrategy=Vanilla ToT, Inference Budget=C=162026.03 | 84.5 | |
| Best-of-16Strategy=Best-of-16, Inference Budget=C=162026.03 | 82.2 | |
| Greedy CoTStrategy=Greedy CoT, Inference Budget=C=162026.03 | 78.5 | |
| BaselineCondition=Shared unadapted baseline2026.06 | 70.74 | |
| ACECondition=Prompt learning generalized from main four-benchmark suite2026.06 | 69.32 | |
| EeveeCondition=Prompt learning generalized from main four-benchmark suite2026.06 | 68.92 | |
| GEPACondition=Prompt learning generalized from main four-benchmark suite2026.06 | 68.85 |