Daniel & Friends / Glossary / Chain of thought
AI and automationChain of thought
Chain of thought is an approach where a model writes out intermediate steps before the final answer. For multi-step tasks such as calculations or logic problems, it usually improves accuracy. Reasoning models carry out this phase on their own.
Daniel Votruba ·
In practice
With standard models you can ask for step-by-step work in the prompt. With reasoning models it is usually unnecessary, so let them think their own way.
Related terms
- Reasoning model: A reasoning model is a language model that spends time thinking before it answers: it breaks a task into steps, tries options and checks itself.
- Prompt engineering: Prompt engineering is the design and testing of instructions for a model: choosing the role, context, examples and answer format so that the output is reliably usable.
- Few-shot and zero-shot prompting: Zero-shot prompting gives a model a task without a single example, while few-shot adds several sample pairs of input and output to the prompt.
- Large language model (LLM): A large language model is a model trained on a huge amount of text that predicts the next words and can write, summarise and answer.
More terms in the group AI and automation
ChatGPT Ads · C2PA and content provenance (AI watermarks) · Computer vision · Automation platforms (Zapier, Make, n8n) · Context window · AI workflow and orchestration · Copilot (AI assistant) · AI slop · Custom GPTs and projects · AI search (answer engines) · Deepfake and synthetic media · AI SDR · Deep learning · AI scoring · Diffusion models · AI safety and alignment · Embedding · AI personalization · EU AI Act · AI governance · Explainable AI (XAI) · AI crawlers (GPTBot and others) · AI costs and limits · Fine-tuning
Terms in practice
Need help putting them to work?
Terms are one thing, putting them to work is another. We go through which of them pay off for you and say so plainly. You get a first opinion for free and we reply within 24 hours.
Book a call