Training Verifiers to Solve Math Word Problems

TL;DR

This paper introduces GSM8K, a diverse dataset of math word problems, and proposes training verifiers to improve model accuracy in solving these problems, demonstrating that verification enhances performance more effectively than finetuning.

Contribution

The paper presents a new dataset GSM8K and a verification-based approach to improve math problem-solving accuracy in language models, showing verification's scalability over finetuning.

Findings

Verification improves accuracy on GSM8K.

Verification scales better with more data than finetuning.

Large models still struggle with multi-step math reasoning.

Abstract

State-of-the-art language models can match human performance on many tasks, but they still struggle to robustly perform multi-step mathematical reasoning. To diagnose the failures of current models and support research, we introduce GSM8K, a dataset of 8.5K high quality linguistically diverse grade school math word problems. We find that even the largest transformer models fail to achieve high test performance, despite the conceptual simplicity of this problem distribution. To increase performance, we propose training verifiers to judge the correctness of model completions. At test time, we generate many candidate solutions and select the one ranked highest by the verifier. We demonstrate that verification significantly improves performance on GSM8K, and we provide strong empirical evidence that verification scales more effectively with increased data than a finetuning baseline.