Glossary
Embedding — Glossary
A numeric vector representation of an item—such as a token, sentence, image, or document—used to capture similarity and feed neural models.
Full definition
Embeddings place discrete objects into a continuous vector space where geometric relationships encode semantic or task-relevant similarity. Word and sentence embeddings support search, clustering, recommendation, and classification. In deep networks, embeddings are learned layers that map tokens or other IDs to vectors.
Quality depends on training objectives and domain. Embedding models used for retrieval should be evaluated with realistic queries and corpora, not only with generic similarity demos.