Cognition · Data artifact
Chain-of-Thought Prompt
Data artifactCognitionCognition & MemoryVariation point (abstract)arc:ChainOfThoughtPrompt
A prompt artifact that elicits explicit, step-by-step intermediate reasoning from a model before it states its final answer, with or without worked demonstrations.
Responsibility. Elicits explicit intermediate reasoning steps from the model.
Also known as: CoT prompt, Reasoning prompt, Visible reasoning chain
Variants
| Variant | When to choose |
|---|---|
| Auto-CoT Prompt | Choose when many distinct problem types need structured reasoning but resources are insufficient to hand-craft examples for each, or when the question distribution is unknown or evolving; accept added pipeline complexity and slightly lower quality than hand-crafted few-shot. |
| Few-Shot CoT Prompt | Choose when reasoning structure and consistency matter more than deployment speed (e.g., medical diagnosis, legal analysis, financial modelling) and expert time can be invested in two to three exemplary demonstrations per problem type. |
| Zero-Shot CoT Prompt | Choose when immediate deployment is needed across diverse, unpredictable problem types without time or budget to craft examples, accepting variable reasoning quality on ambiguous problems. |
Relationships
configures structural
is evaluated by assurance
Design guidance
- SHOULD start with zero-shot CoT for prototyping and simple domains and invest in demonstrations only when quality inconsistencies or missing domain reasoning patterns are observed.
- SHOULD validate CoT benefit against direct prompting on the target problem distribution, since multi-hop tasks with long chains can perform worse with CoT.
- SHOULD NOT apply full CoT integration to simple lookups or fixed procedural tasks where reasoning quality cannot be validated from outcomes.
Quantitative guidance
As stated by the sources; verify before use.
- Few-shot CoT reported accuracy improvements up to 28.2% on certain reasoning tasks versus zero-shot (Ch5.1).
Classification
- Patterns
- Chain-of-ThoughtZero-shot CoTFew-shot CoTAuto-CoTExternalized reasoning for user transparency
- Quality attributes
- Transparency and accountability (NIST AI RMF: accountable and transparent)Functional suitability: correctness and validity (ISO/IEC 25010 | NIST AI RMF: valid)Maintainability (ISO/IEC 25010)
- Risks mitigated
- Opaque direct answers that cannot be validatedMulti-hop reasoning errors from single-step answering
Sources
- Ch5.1: T. Nguyen, "Chain-of-Thought (CoT) Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.1. ISBN: 9798244538229.
- Ch5.3: T. Nguyen, "Self-Consistency Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.3. ISBN: 9798244538229.
- Ch10.1: T. Nguyen, "Conversational UI," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.1. ISBN: 9798244538229.
- Ch10.2: T. Nguyen, "Proactive Agents," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.2. ISBN: 9798244538229.
- Ch10.4: T. Nguyen, "Human-in-the-Loop," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.4. ISBN: 9798244538229.