Reasoning on BBH (Score, Delta, Avg. Rank)
80.1BBH ScoreQwen2.5(14B)T
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen2.5(14B)TEvaluation Protocol=Generation-based2026.07 | 80.1 | — | — | |
| Qwen2.5(14B)TEvaluation Protocol=Likelihood-based2026.07 | 51.71 | — | — | |
| Qwen2.5(1.5B)SEvaluation Protocol=Generation-based2026.07 | 47.09 | 0 | 2.4 | |
| Reverse KLEvaluation Protocol=Generation-based2026.07 | 44.52 | -2.57 | 3 | |
| RG-OPDEvaluation Protocol=Generation-based2026.07 | 44.04 | -3.05 | 2 | |
| TSD-KDEvaluation Protocol=Likelihood-based2026.07 | 37.96 | 1.77 | 2.6 | |
| TSD-KDEvaluation Protocol=Generation-based2026.07 | 37.08 | -10.01 | 2.6 | |
| Qwen2.5(1.5B)SEvaluation Protocol=Likelihood-based2026.07 | 36.19 | 0 | 2.4 | |
| RG-OPDEvaluation Protocol=Likelihood-based2026.07 | 34.4 | -1.79 | 2 | |
| Reverse KLEvaluation Protocol=Likelihood-based2026.07 | 32.7 | -3.49 | 3 |