DictionaryArchitecture
Embedding
Turning text into a list of numbers that captures its meaning, so a computer can find similar ideas, not just matching words.
Definition
A list of numbers (a vector) that represents the meaning of a piece of text, so similar texts sit close together and can be compared with cosine or dot-product similarity. Embeddings power semantic search and RAG. Anthropic does not offer its own embedding model; its docs point to providers such as Voyage AI, whose current general models return 1024-dimensional vectors by default.
Example
'Reset my password' and 'I cannot log in' share few words, but their embeddings are close, so search returns the same help article.