Embedding

An Embedding is a way of translating text (or images/audio) into a list of numbers (a vector) that captures its semantic meaning.

While humans understand words, computers understand numbers. Embeddings allow AI models to determine that "cat" and "kitten" are similar concepts (close together in vector space), whereas "cat" and "car" are different, even if they share similar letters. This technology underpins semantic search and RAG.