Towards Reasoning in Large Language Models: A Survey

TL;DR

This survey reviews the current state of reasoning capabilities in large language models, discussing techniques, evaluation methods, findings, and future research directions to understand and enhance their reasoning abilities.

Contribution

It provides a comprehensive overview of techniques, benchmarks, and research findings related to reasoning in large language models, highlighting gaps and future directions.

Findings

LLMs show emerging reasoning abilities with increased size

Existing benchmarks reveal limitations in LLM reasoning

Techniques for improving reasoning are actively being developed

Abstract

Reasoning is a fundamental aspect of human intelligence that plays a crucial role in activities such as problem solving, decision making, and critical thinking. In recent years, large language models (LLMs) have made significant progress in natural language processing, and there is observation that these models may exhibit reasoning abilities when they are sufficiently large. However, it is not yet clear to what extent LLMs are capable of reasoning. This paper provides a comprehensive overview of the current state of knowledge on reasoning in LLMs, including techniques for improving and eliciting reasoning in these models, methods and benchmarks for evaluating reasoning abilities, findings and implications of previous research in this field, and suggestions on future directions. Our aim is to provide a detailed and up-to-date review of this topic and stimulate meaningful discussion and future work.