We introduce SELF-DISCOVER, a general framework for LLMs to self-discover the task-intrinsic reasoning structures to tackle complex reasoning problems that are challenging for typical prompting methods. Core to the framework is a self-discovery process where LLMs select multiple atomic reasoning modules such as critical thinking and step-by-step thinking, and compose them into an explicit reasoning structure for LLMs to follow during decoding. SELF-DISCOVER substantially improves GPT-4 and PaLM 2's performance on challenging reasoning benchmarks such as BigBench-Hard, grounded agent reasoning, and MATH, by as much as 32% compared to Chain of Thought (CoT). Furthermore, SELF-DISCOVER outperforms inference-intensive methods such as CoT-Self-Consistency by more than 20%, while requiring 10-40x fewer inference compute. Finally, we show that the self-discovered reasoning structures are universally applicable across model families: from PaLM 2-L to GPT-4, and from GPT-4 to Llama2, and share commonalities with human reasoning patterns.
@article{arxiv.2402.03620,
title = {Self-Discover: Large Language Models Self-Compose Reasoning Structures},
author = {Pei Zhou and Jay Pujara and Xiang Ren and Xinyun Chen and Heng-Tze Cheng and Quoc V. Le and Ed H. Chi and Denny Zhou and Swaroop Mishra and Huaixiu Steven Zheng},
journal= {arXiv preprint arXiv:2402.03620},
year = {2024}
}