Chain-of-Thought (CoT)
Also known as: CoT, Chain of thought, Step-by-step reasoning
Ask the model to think step by step before answering. It boosts accuracy on reasoning tasks.
A prompting technique that asks the model to produce intermediate reasoning steps before the final answer. Improves performance on multi-step reasoning tasks; closely related to how 'reasoning models' like o1 and DeepSeek-R1 are trained.
In practice
Foundational prompting pattern. Modern reasoning models bake CoT into their RL training, so understanding it explains o1/R1's behavior.
How it compares
CoT is a specific prompting technique; prompt engineering is the broader practice.
Comparisons that include Chain-of-Thought (CoT)
Related topics
Questions that mention this term
- Explain G-Eval's mechanism and why it improves over direct score LLM judges
- How does a chain-of-thought judge differ from a direct score judge, and when is each better?
- What is G-Eval and how does it differ from a direct score LLM judge?
- Challenge this claim that reasoning models always expose full chain-of-thought
- What is the primary advantage of chain-of-thought (CoT) judge prompts over direct score prompts?
- Flashcard: what is chain-of-thought (CoT) prompting?
Related terms
Agent Loop
The runtime: LLM call → tool call → observation → LLM call → … until a final answer or step cap.
AI Agents
LLMs that loop: plan → call tools → observe results → repeat until done.
Context Rot
Long-running chats degrade: early instructions get forgotten, and tool calls become less reliable.
Function Calling
The model emits structured JSON specifying a tool call instead of free text; the host executes it.
In-Context Learning (ICL)
Show the model a few examples in the prompt and it learns the pattern, no fine-tuning needed.
Model Context Protocol (MCP)
An open protocol (from Anthropic) for letting LLMs call tools and read resources from any compatible server.