Chain-of-Thought Prompting
A prompting technique that encourages a language model to break down complex reasoning tasks into intermediate steps, improving accuracy on logical and mathematical problems.
In plain language
Asking an AI to think step-by-step and show its reasoning before answering. This simple instruction often dramatically improves accuracy on maths and logic problems, similar to asking someone to show their working on an exam.
Why this matters
Chain-of-thought prompting improves both accuracy and interpretability of language model outputs. For organisations deploying language models in decision support roles, requiring models to show reasoning enhances user trust and enables humans to catch errors in the AI's logic.
Relevance
ImplementationChain-of-thought is a prompting technique that improves model interpretability and accuracy for decision support applications.
Related terms
Putting chain-of-thought prompting into practice in your organisation?
Ready to transform your AI strategy?
Partner with Australia's AI strategy and governance specialists. From adoption roadmaps to ISO 42001 audit readiness.