Foundations
Retrieval-Augmented Generation (RAG)
Also known as RAG
Retrieval-Augmented Generation (RAG) retrieves relevant passages at query time and feeds them into an LLM's context.
It answers from your own data instead of the model's training memory, trading a knowledge problem for a retrieval-quality one.
Chunking, embeddings and ranking decide whether the answer is grounded or just a plausible-sounding guess.
Most complaints that the AI does not know your data point to a RAG pipeline defect, not a model limit.
← Back to the full glossaryRelated terms
Embeddings Embeddings are numeric vectors that place similar content close together in vector space. Vector Search (Semantic Search) Vector search finds the nearest matches to a query by comparing embeddings, not exact keywords. Chunking Chunking splits source documents into smaller passages before embedding, so retrieval returns focused text. Hallucination A hallucination is a confident, fluent output that is factually wrong or unsupported.
More in Foundations
Context Window The context window is the maximum text, measured in tokens, a model can consider at once. Token A token is the basic unit of text an LLM reads and writes. Model-Agnostic Architecture A model-agnostic architecture lets an application swap or route between LLM providers without a rewrite. Fine-Tuning (vs RAG) Fine-tuning trains a model's weights on your own examples, changing its behavior directly.