KNOWLEDGE, EMBEDDINGS & RAG • 142 / 397
Teach models to look up relevant information before answering.
Retrieval-Augmented Generation (RAG)
RAG retrieves relevant external information and gives it to the model before generation.
Think of it like
Think of Retrieval-Augmented Generation (RAG) like a library system: instead of memorizing every book, find the right shelf and bring only the useful pages to the desk.
Real life
Search engines and recommendation systems use ideas related to Retrieval-Augmented Generation (RAG).
SRE lens
Ask “have we seen this DB timeout before?” and retrieve previous incidents before answering.
Remember thisRAG retrieves relevant external information and gives it to the model before generation.