training
RAG
Retrieval-Augmented Generation, a technique that enhances language model outputs by retrieving relevant information from external knowledge sources before generating a response. RAG reduces hallucinations and keeps responses current without retraining the model.
In practice
A RAG-powered legal assistant retrieves relevant case law from a database before answering questions about legal precedent.
In the index
Tools that mention RAG
Matched on each tool’s own description and feature list, highest trust score first.