RAG (Retrieval-Augmented Generation)
Enhancing model responses by fetching relevant documents from an external knowledge base at query time.
RAG combines a retrieval system (typically a vector database) with a generative model. When a query arrives, relevant passages are fetched and injected into the prompt context before the model responds. This keeps responses factually grounded without the cost of fine-tuning, allows knowledge to be updated without retraining, and is the dominant pattern for production knowledge-base chatbots.
Termes Associés
A database optimized for storing and searching high-dimensional embedding vectors.
A dense numerical vector that encodes the semantic meaning of text.
Connecting model responses to verified, real-world information sources.
When a model generates plausible-sounding but factually incorrect information.