Embeddings are numerical representations of text, images, or other data that capture semantic meaning as position in a mathematical space, rather than encoding words as arbitrary, unrelated symbols the way a simple token ID would. Text with similar meaning ends up with similar embeddings, which lets a model measure how closely related two pieces of content actually are.
This numerical, meaning-aware representation is what makes many practical AI applications possible: semantic search uses embeddings to find documents related to a query even when they don't share exact keywords, recommendation systems use them to find similar items, and retrieval-augmented generation systems use them to pull relevant context into a language model's prompt before it generates a response.
