What are embeddings?

An embedding is a list of numbers that represents the meaning of a piece of text, produced by an embedding model. Texts with similar meaning get embeddings that are close together, so comparing embeddings finds related passages even when they share no words. Embeddings power semantic search, recommendations and retrieval-augmented generation.

A typical text embedding has a few hundred to a few thousand numbers. Closeness is usually measured with cosine similarity.

Embedding models are much smaller than chat models and can run on a laptop, which makes private search possible.