Fast, Slow, and Tool-augmented Thinking for LLMs: A Review

TL;DR

This paper reviews various reasoning strategies in Large Language Models, proposing a taxonomy based on speed and tool use, and surveys recent methods to enhance adaptive reasoning capabilities.

Contribution

It introduces a novel taxonomy of LLM reasoning strategies along two boundaries and systematically categorizes recent adaptive reasoning methods in LLMs.

Findings

Identification of fast/slow and internal/external reasoning boundaries

Survey of recent adaptive reasoning techniques in LLMs

Highlighting open challenges and future research directions

Abstract

Large Language Models (LLMs) have demonstrated remarkable progress in reasoning across diverse domains. However, effective reasoning in real-world tasks requires adapting the reasoning strategy to the demands of the problem, ranging from fast, intuitive responses to deliberate, step-by-step reasoning and tool-augmented thinking. Drawing inspiration from cognitive psychology, we propose a novel taxonomy of LLM reasoning strategies along two knowledge boundaries: a fast/slow boundary separating intuitive from deliberative processes, and an internal/external boundary distinguishing reasoning grounded in the model's parameters from reasoning augmented by external tools. We systematically survey recent work on adaptive reasoning in LLMs and categorize methods based on key decision factors. We conclude by highlighting open challenges and future directions toward more adaptive, efficient, and reliable LLMs.