Chain of Thought

Definition

A prompting technique where you ask an AI model to show its reasoning step by step before giving a final answer. Dramatically improves performance on math, logic, and complex reasoning tasks. Models that 'think out loud' make fewer errors because each step can build on verified intermediate results.

How It Works

Chain-of-thought prompting is a technique where you ask an AI model to show its reasoning step by step before giving a final answer. Instead of jumping straight to a conclusion, the model works through the problem piece by piece. It turns out this simple trick dramatically improves performance on math, logic, and complex reasoning tasks.

The discovery was surprisingly recent — a 2022 paper from Google showed that just adding "Let's think step by step" to a prompt could boost accuracy on math problems from around 18% to 79% with the same model. The model already had the capability; it just needed permission (and space) to reason out loud. This sparked an entire research direction around eliciting better reasoning from existing models.

Chain-of-thought has evolved into more sophisticated techniques like tree-of-thought (exploring multiple reasoning paths) and self-consistency (generating several reasoning chains and picking the most common answer). Modern models like Claude and GPT-4 are trained to reason step by step by default for complex queries, which is why their responses often walk through logic before stating conclusions.