Forward-Backward Reasoning in Large Language Models for Mathematical Verification

arXiv:2308.07758v5 Announce Type: replace-cross
Abstract: Self-Consistency samples diverse reasoning chains with answers and chooses the final answer by majority voting. It is based on forward reasoning and cannot further improve performance by sampling more reasoning chains when saturated. To further boost performance, we introduce backward reasoning to verify candidate answers. Specifically, for mathematical tasks, we mask a number in the question and ask the LLM to answer a backward question created by a simple template, i.e., to predict the masked …

abstract arxiv boost cs.ai cs.cl cs.lg diverse language language models large language large language models performance reasoning samples sampling type verification verify voting