What are embeddings? Search by meaning example.

Embeddings

Dense vectors represent meaning—similar sentences have close cosine distance. Enables semantic search over docs.

similarity = cosine(embed("refund policy"), embed("how do I get money back"))