Chain of Thought.
A prompting technique that encourages an LLM to reason step by step before giving a final answer, dramatically improving performance on complex tasks.
Chain of Thought (CoT) prompting was popularized by the 2022 Google paper 'Chain-of-Thought Prompting Elicits Reasoning in Large Language Models.' The core insight is that asking a model to 'think step by step' causes it to generate intermediate reasoning traces that dramatically improve accuracy on multi-step problems like math, logic, and science questions.
CoT works because the intermediate tokens serve as a scratchpad — each step in the reasoning creates context for the next step, allowing the model to tackle problems that exceed its ability to solve in a single forward pass. The technique essentially lets the model allocate more compute to hard problems by spending more tokens thinking about them.
Modern reasoning models like OpenAI's o-series and Anthropic's Claude 3.7 Sonnet with extended thinking take this further, performing extended internal CoT before generating a response. These models learn through reinforcement learning to produce useful reasoning traces, rather than relying on prompting alone.
Related terms.
- RLHFReinforcement Learning from Human Feedback — a training technique that uses human preferences to teach AI models to be helpful, honest, and harmless.
- Prompt EngineeringThe practice of crafting inputs, instructions, examples, and formatting to guide language models toward producing better, more accurate outputs.