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

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zheng, Huaixiu Steven, Mishra, Swaroop, Chen, Xinyun, Cheng, Heng-Tze, Chi, Ed H., Le, Quoc V, Zhou, Denny
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911793213865984
author Zheng, Huaixiu Steven
Mishra, Swaroop
Chen, Xinyun
Cheng, Heng-Tze
Chi, Ed H.
Le, Quoc V
Zhou, Denny
author_facet Zheng, Huaixiu Steven
Mishra, Swaroop
Chen, Xinyun
Cheng, Heng-Tze
Chi, Ed H.
Le, Quoc V
Zhou, Denny
contents 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%.
format Preprint
id arxiv_https___arxiv_org_abs_2310_06117
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Take a Step Back: Evoking Reasoning via Abstraction in Large Language Models
Zheng, Huaixiu Steven
Mishra, Swaroop
Chen, Xinyun
Cheng, Heng-Tze
Chi, Ed H.
Le, Quoc V
Zhou, Denny
Machine Learning
Artificial Intelligence
Computation and Language
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%.
title Take a Step Back: Evoking Reasoning via Abstraction in Large Language Models
topic Machine Learning
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2310.06117