An embedding is a vector (a list of numbers) that represents a piece of text (a token, sentence, or document) as a point in high-dimensional space. The defining property: semantically similar things land close together, so meaning becomes distance.
Embeddings power semantic search, clustering, recommendation, and retrieval-augmented generation, where a vector database finds the nearest neighbours to a query. They’re also the input to attention inside a transformer.
Play with a 2-D map of embeddings in Embeddings, explained.