Enhancing Chain-of-Thoughts Prompting with Iterative Bootstrapping in Large Language Models

arXiv:2304.11657v3 Announce Type: replace
Abstract: Large language models (LLMs) can achieve highly effective performance on various reasoning tasks by incorporating step-by-step chain-of-thought (CoT) prompting as demonstrations. However, the reasoning chains of demonstrations generated by LLMs are prone to errors, which can subsequently lead to incorrect reasoning during inference. Furthermore, inappropriate exemplars (overly simplistic or complex), can affect overall performance among varying levels of difficulty. We introduce Iter-CoT (Iterative bootstrapping in Chain-of-Thoughts Prompting), an iterative bootstrapping approach for selecting exemplars and …

abstract arxiv bootstrapping cs.ai cs.cl errors generated however inappropriate inference iterative language language models large language large language models llms performance prompting reasoning step-by-step tasks thought thoughts type