AIIntermediate#AI Safety#Evidence#Prompting
Chain of Thought (CoT)
C
Definition
A prompting pattern that asks a model to decompose a task into intermediate text. The text may help a person inspect a draft, but it is generated output, not proof of the model's internal process, factual accuracy, or a valid financial conclusion.
Source reviewed
- NIST AI RMF: Generative AI Profile opens in a new tab
Supports: NIST explains that generative systems can confidently produce false or internally inconsistent content, including fabricated logic or citations that can mislead people who act on the output.
- OWASP: LLM06 Excessive Agency opens in a new tab
Supports: OWASP explains that unexpected, ambiguous, or manipulated LLM output can cause damaging actions when an application grants broad authority to tools.
Concept Diagram
Key Takeaways
1
CoT is a generated intermediate-text pattern; it is not evidence of a model's internal process or of a correct conclusion.
2
A reviewable explanation should separate source records, observed facts, assumptions, reproducible calculations, missing data, and tentative summary.
3
Evaluate claims against primary records and recompute key calculations instead of trusting a fluent step-by-step narrative.
4
Generated steps and conclusions must remain drafts: state-changing actions require independent policy checks and explicit approval.
Practical Example
A model receives two dated public reports about a protocol event. It lists each source, extracts only attributed statements, notes an unresolved contradiction, and produces questions for a reviewer. It does not infer a price direction, label the explanation an audit trail, or prepare a trade.
Finished learning?
Related Terms
AI
Perception LayerThe part of an AI workflow that collects and labels inputs such as blockchain records, market data, and public communications. It supplies context for review; it does not establish truth, identify people, predict prices, or justify an action.
AI
LLM ReasoningUsing a language model to summarize, classify, or compare supplied information. Its output is probabilistic text, not verified market analysis, financial advice, or authority to execute a trade.
AI
Prompt EngineeringThe practice of designing model instructions, examples, context, and output constraints for a bounded task. It can make a workflow easier to evaluate; it does not make model output deterministic, correct, secure, or suitable for a financial decision.
AI
RAG (Retrieval-Augmented Generation)A system pattern that retrieves selected external records and supplies them with a query to a language model. Retrieval can improve traceability and recency when sources are cited; it does not prove that a record is true, complete, current, or safe to act on.
Related Deep Dives
Master the hub of AI
Explore all our strategic guides about AI to take your operations to the next level.
View all articles