Retrieval-Augmented Generation (RAG)
概念
训练与推理
Look it up, then answer — how enterprises ground models in their data.
您正在查看一条概念解释记录：该术语的通俗含义，以及它所关联的各条目。
RAG retrieves relevant documents (via embedding search) and feeds them into the model's context, giving current, citable, private knowledge without retraining. It is the default enterprise AI architecture and the reason vector databases and embedding models became infrastructure categories; long context windows compete with it, agents combine both.
它如何与整个技术栈相连
本记录的关联链接一览——点击分组可直接跳转。
来源
记录审阅截至 2026-01-15 — 查看 方法论.