AI Glossary

Retrieval-augmented generation

Also called RAG

What does “Retrieval-augmented generation” mean?

Retrieval-augmented generation: Pattern where a system first retrieves relevant information, such as documents, database records or search results, and adds it to the prompt so the model can answer using that material. It lets a model work with private or recent information without retraining and makes answers easier to check against sources.

Retrieval-augmented generation visual guide: Pattern where a system first retrieves relevant information, such as documents, database records or search results, and adds it to the prompt so the model can answer using that material. It lets a model work with private or recent information without retraining and makes answers easier to check against sources.
Visual guide Open full size (new tab)

What is an example of “Retrieval-augmented generation”?

An internal HR assistant uses RAG to pull the current leave policy and quote it, instead of guessing from general knowledge.

Keep learning

See also

EmbeddingsTermVector databaseTermChunkingTermGroundingTerm

Same topic

Agent memoryTermContext windowTermEmbedding modelTermHallucinationTermKnowledge cutoffTerm