Take a Step Back: Evoking Reasoning via Abstraction in Large Language Models
This addresses the challenge of enhancing reasoning abilities in LLMs for tasks like STEM and QA, representing a novel method for a known bottleneck.
The paper tackles the problem of improving reasoning in large language models by introducing Step-Back Prompting, a technique that uses abstraction to derive high-level concepts and first principles, resulting in substantial performance gains such as 7-11% improvements on MMLU and 27% on TimeQA.
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 Multi-Hop Reasoning. For instance, Step-Back Prompting improves PaLM-2L performance on MMLU (Physics and Chemistry) by 7% and 11% respectively, TimeQA by 27%, and MuSiQue by 7%.