OVM, Outcome-supervised Value Models for Planning in Mathematical Reasoning

Benchmark Model Rank Results
arithmetic-reasoning-on-gsm8kOVM-Mistral-7B (verify100@1)#119Accuracy: 84.7Parameters (Billion): 7
arithmetic-reasoning-on-gsm8kOVM-Mistral-7B (verify20@1)#140Accuracy: 82.6Parameters (Billion): 7
arithmetic-reasoning-on-gsm8kOVM-Llama2-7B (verify100@1)#191Accuracy: 73.7Parameters (Billion): 7