# Parametric knowledge

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

Also called: trained knowledge, built-in knowledge, model knowledge, world knowledge

What a model knows from its training, stored in its [parameters](https://ecosystem.samcarlton.com/parameters). It is broad but frozen at the [knowledge cutoff](https://ecosystem.samcarlton.com/knowledge-cutoff), and the model cannot say exactly where any given fact came from.

**For example:** A model can explain what a [SAFE](https://ecosystem.samcarlton.com/safe) is from parametric knowledge, but it cannot know about a local [grant](https://ecosystem.samcarlton.com/grant) program announced last month.

## Connects to (6)

- [Contextual knowledge](https://ecosystem.samcarlton.com/contextual-knowledge): Facts the model can read directly in front of it right now, such as documents you pasted, files the agent opened or search results.
- [Parameters](https://ecosystem.samcarlton.com/parameters): The values a model learns during training.
- [Training](https://ecosystem.samcarlton.com/training): Teaching an AI model by showing it a very large number of examples before anyone uses it.
- [Knowledge cutoff](https://ecosystem.samcarlton.com/knowledge-cutoff): The date where a model's training data ends.
- [Hallucination](https://ecosystem.samcarlton.com/hallucination): When an AI model states something false or invented as though it were true.
- [Weights](https://ecosystem.samcarlton.com/weights): The numbers inside a model that it learned during training.

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