AI Agent Architecture — ReAct, Memory, Planning and Multi-Agent Systems · Reasoning — How an LLM Thinks Before Acting

Zero-shot CoT vs few-shot CoT — When Each Works

Reasoning — How an LLM Thinks Before Acting

Introduction

Kojima et al. (2022) discovered that adding the phrase "Let's think step by step" to a prompt with no exemplars (zero-shot CoT) already triggers chain reasoning in large models. This lesson compares zero-shot CoT with few-shot CoT in terms of gain, cost, and use cases — when a single magic prompt suffices and when carefully chosen examples are necessary.