List your company

Mathematical Reasoning on GSM8K (Accuracy & Max Accuracy)

58.4AccuracyReSCALE
53.254.5555.957.25Mar 22, 2026

Evaluation results

MethodLinks
ReSCALEBudget=Large, Tokens=16K–18K, Backbone=Llama2-7B2026.0358.459.7
ReSCALEBudget=Medium, Tokens=6K–8K, Backbone=Llama2-7B2026.0357.960.2
AlphaZeroBudget=Medium, Tokens=6K–8K, Backbone=Llama2-7B2026.0356.758.3
AlphaZeroBudget=Small, Tokens=0.5K–2K, Backbone=Llama2-7B2026.035658.8
ReSCALEBudget=Small, Tokens=0.5K–2K, Backbone=Llama2-7B2026.0355.158.6
AlphaZeroBudget=Large, Tokens=16K–18K, Backbone=Llama2-7B2026.0353.655
Best-of-NBudget=N = 32, Tokens=3.5K, Backbone=Llama2-7B2026.0353.4—