Embedding
A vector representation of text such that semantically similar inputs land near each other.
By Lattice editorial · Entry last reviewed · method
What it implies
The same model must produce embeddings for documents and queries, or the space is meaningless. The failure mode to know about: embeddings are bad at exact tokens — part numbers, error codes, names — because `ERR-4421` and `ERR-4412` really are nearly the same thing. Users type exact tokens constantly.
Related terms
- Hybrid searchCombining lexical retrieval (BM25) with vector retrieval, so exact matches and semantic matches are both found.
- Vector indexAn approximate structure — usually HNSW or an IVF variant — that finds nearest neighbours without scanning every vector.
- ChunkingSplitting documents into retrievable pieces, which is where RAG quality is mostly won and lost.