Chain-of-thought
A prompting and training technique where a model works through intermediate steps in writing before giving a final answer, often improving accuracy on multi-step problems.
Chain-of-thought is the practice of getting a model to spell out intermediate reasoning steps rather than jumping straight to an answer. On many maths, logic, and coding tasks, thinking “out loud” in this way measurably improves results.
The idea started as a prompting trick (“let’s think step by step”) and later became something models are trained to do. Reasoning models such as OpenAI o1 and DeepSeek-R1 use reinforcement learning to produce long internal chains before responding, spending more computation at answer time to get a better result.