Retrieval-Augmented Generation (RAG)

Retrieval-Augmented Generation (RAG) combines external information retrieval with large language models to produce more factual and up-to-date responses. The concept integrates search, indexing and reranking components with generative models and defines interfaces, evaluation criteria and governance for knowledge-intensive applications. RAG addresses answer…

Use this profile to understand the building block briefly, place it in the model, and switch to the 360° assessment when needed.

Content type

Concept

Theoretical construct: explains a term, principle, or mental model.

Classification level

Foundation

What you need to understand to reason about a domain.

Available perspectives

Open this building block directly in an already curated view.

360°

In brief

Retrieval-Augmented Generation in brief

Retrieval-augmented generation (RAG) supplements a language model at response time with relevant information retrieved from an external knowledge source.

Where does it come from?

Patrick Lewis and co-authors described RAG as a research approach in 2020. It combines information retrieval with generative modeling so knowledge outside model weights can remain usable and updateable.

The mental model

A query is interpreted, searched, and ranked; selected passages are provided as context to the model, which generates the answer. Quality therefore depends on chunking, the index, retrieval, context window, prompt, model, and evaluation. Citations help but do not replace checking the data and answer.

Key building blocks

Retriever

A search component finds and ranks documents or passages for a query.

Grounding context

Selected sources constrain and inform generation instead of relying only on model parameters.

RAG evaluation

Retrieval and answer quality are measured separately and together with real questions and error classes.

When is it useful?

RAG fits changing, domain-specific, or private knowledge bases. It does not automatically remove hallucinations or retrieval errors; access control, freshness, source quality, and cost belong in operations.

Lewis et al. – Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks Microsoft Learn – Retrieval-augmented generation