# Contextual knowledge

Concept · [Building with AI](https://ecosystem.samcarlton.com/ai)

Also called: in-context knowledge, context knowledge, grounding, grounded answer

Facts the model can read directly in front of it right now, such as documents you pasted, files the [agent](https://ecosystem.samcarlton.com/ai-agent) opened or search results. This information is current and checkable, which is why loading the right material is the main defense against [made-up answers](https://ecosystem.samcarlton.com/hallucination).

**For example:** Instead of asking the model from memory, a [founder](https://ecosystem.samcarlton.com/founder) pastes in this year's program rules, so the answer comes from contextual knowledge.

## Connects to (7)

- [Parametric knowledge](https://ecosystem.samcarlton.com/parametric-knowledge): What a model knows from its training, stored in its parameters.
- [Context window](https://ecosystem.samcarlton.com/context-window): How much text an LLM can take into account at one time.
- [RAG](https://ecosystem.samcarlton.com/rag): Retrieval-augmented generation.
- [Tool result](https://ecosystem.samcarlton.com/tool-result): What comes back to the model after it uses a tool: a file's contents, the output of a command, a search result or an error.
- [Hallucination](https://ecosystem.samcarlton.com/hallucination): When an AI model states something false or invented as though it were true.
- [Knowledge cutoff](https://ecosystem.samcarlton.com/knowledge-cutoff): The date where a model's training data ends.
- [Primary source](https://ecosystem.samcarlton.com/primary-source): The original material itself, such as the code, a full transcript or the raw data.

Map view: https://ecosystem.samcarlton.com/contextual-knowledge