Definition
An embedding model turns content into a vector, a list of hundreds or thousands of numbers. Content with similar meaning produces vectors that are close to each other, which makes it possible to search by meaning, group similar items, recommend related content and detect duplicates.
Embeddings power semantic search and the retrieval step of RAG. They are cheap to compute compared with generating text, but they must be recomputed when you change the embedding model, so the choice of model is a decision worth making carefully.