List your company

Mathematical Reasoning on GSM8K (Mono/Non-Mono Accuracy & Gap)

73.7Monotonicity RateQwen2.5-7B-Instruct
70.01571.857573.775.5425Mar 19, 2026

Evaluation results

MethodLinks
Qwen2.5-7B-Instructn=300, m=5, τ=0.72026.0373.768.846.921.9