Mathematical Reasoning on GSM8K n=300 (test)
73.7Monotonicity Rateentropy-trajectory monotonicity
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| entropy-trajectory monotonicityModel=Qwen2.5-7B-Instruct, n=300, m=5, tau=0.7, seed=422026.03 | 73.7 | 68.8 | 46.8 | 21.9 | |
| entropy-trajectory monotonicityModel=Mistral-7B-Instruct-v0.3, n=300, m=5, tau=0.7, seed=422026.03 | 39.7 | 72.3 | 37.6 | 34.7 |