Retrieval-Augmented Generation
Retrieval-Augmented Generation is an architecture that enriches a language model's input with documents fetched from an external corpus at query time. A retriever ranks passages by semantic similarity, and the top results are injected into the prompt so the model can produce answers grounded in source material rather than relying solely on parameters learned during training.
Definition
Retrieval-Augmented Generation is an architecture that enriches a language model's input with documents fetched from an external corpus at query time. A retriever ranks passages by semantic similarity, and the top results are injected into the prompt so the model can produce answers grounded in source material rather than relying solely on parameters learned during training.
Examples
- Customer support bot retrieves product manuals before answering a refund policy question.
- Legal assistant pulls case law from a firm's archive to cite precedents in drafts.
- Internal HR chatbot fetches the latest benefits handbook to answer enrollment questions.
Related terms
No credit card required
Frequently asked
- What is Retrieval-Augmented Generation?
- Retrieval-Augmented Generation is an architecture that enriches a language model's input with documents fetched from an external corpus at query time. A retriever ranks passages by semantic similarity, and the top results are injected into the prompt so the model can produce answers grounded in source material rather than relying solely on parameters learned during training.
- How is Retrieval-Augmented Generation used in practice?
- Retrieval-Augmented Generation is commonly used in scenarios such as: Customer support bot retrieves product manuals before answering a refund policy question.; Legal assistant pulls case law from a firm's archive to cite precedents in drafts.; Internal HR chatbot fetches the latest benefits handbook to answer enrollment questions..
- How does yno.ai support Retrieval-Augmented Generation?
- yno.ai's tooling layer makes Retrieval-Augmented Generation available across all supported models without per-provider boilerplate.
- What other concepts are related to Retrieval-Augmented Generation?
- Retrieval-Augmented Generation relates to Embedding, Vector Database, Knowledge Cutoff, Hallucination. Each links to its own glossary page below.