Chain of Thought (CoT) is a prompting and reasoning technique in large language models (LLMs) that encourages the model to generate intermediate reasoning steps before arriving at a final answer. Instead of directly answering a question, the model is guided to "think aloud" — mimicking how humans solve complex problems by breaking them down into smaller, logical steps.

For example, instead of prompting:

What is 27 × 43?

You might prompt:

Let’s think step by step: First, break 27 into 20 and 7…

This structured, step-by-step output improves model accuracy on tasks requiring multi-step reasoning, such as:

• Math word problems
• Logical puzzles
• Commonsense reasoning
• Program synthesis

CoT prompting can be done in two ways:

1. Zero-shot CoT: Simply appending “Let’s think step by step” to the prompt to trigger reasoning behavior.
2. Few-shot CoT: Including examples of step-by-step reasoning in the prompt (also known as CoT demonstrations).

By guiding LLMs to articulate their reasoning, CoT helps reduce errors caused by rushed or shallow predictions. It also enhances transparency — allowing humans to inspect the logic behind a model’s answer.