Glossary

Foundation model — Glossary

A large pretrained model adaptable to many downstream tasks via prompting, tools, or fine-tuning.

Reviewed:

Full definition

The term was popularized by Stanford CRFM’s foundation model report. Foundation models amortize expensive pretraining across many applications, concentrating both capability and systemic risk.

Adaptation surfaces include prompting, RAG, tool use, and parameter-efficient fine-tuning. Governance should cover the entire adaptation chain, not only the base model card.

Related encyclopedia topic

Foundation Models — Foundation models are large pretrained models adaptable to many downstream tasks via prompting, fine-tuning, or tooling—central to today’s AI product ecosystem.

← Back to glossary · Learning hub