04 · How Models Run
Embedding
Meaning as numbers
A list of numbers (a vector) representing the meaning of a token or document so the model can do math on it. Similar meanings land close together. Embeddings are the bridge between language and machine math — and they power semantic search and RAG.
Concrete example
“King” and “queen” end up near each other in number-space; “king” and “bicycle” do not.
Why it matters
The mechanism that lets software “search by meaning” instead of matching keywords.