Take a Step Back: Evoking Reasoning via Abstraction in Large Language Models

arXiv:2310.06117v2 Announce Type: replace
Abstract: We present Step-Back Prompting, a simple prompting technique that enables LLMs to do abstractions to derive high-level concepts and first principles from instances containing specific details. Using the concepts and principles to guide reasoning, LLMs significantly improve their abilities in following a correct reasoning path towards the solution. We conduct experiments of Step-Back Prompting with PaLM-2L, GPT-4 and Llama2-70B models, and observe substantial performance gains on various challenging reasoning-intensive tasks including STEM, Knowledge QA, and …

abstract abstraction abstractions arxiv concepts cs.ai cs.cl cs.lg guide instances language language models large language large language models llms prompting reasoning simple type via