Skip to content
Agent YAP

What is it?

Embedding

A list of numbers (a vector) that represents the meaning of a piece of text, produced by a model. Texts with similar meaning get vectors that sit close together in this numeric space, which lets a system find related content by measuring distance. Embeddings power semantic search and vector databases. A known weakness: embeddings capture "looks similar" but not "is the cause of," so they can return text that seems related but is not actually relevant.

From the AI agent glossary

As the Foundational Modelling glossary puts it

An embedding is the vector of numbers that represents a token inside the model. When the word "dog" enters a model, it is first converted into its embedding, a long list of numbers that encodes its meaning.

The magic of embeddings is that meaning becomes geometry. Words with similar meanings end up with similar embeddings, sitting close together in the space of numbers. "Dog" and "puppy" land near each other; "dog" and "spreadsheet" land far apart. The model learns these positions during training. Embeddings are the bridge that turns language into math.


See it in the Foundational Modelling glossary

Where Embedding is explained

  1. Retrieval: RAG vs Tools vs Long ContextArchitecture and System Design · The three strategies
  2. How Memory Is Built and RetrievedAgentic Memory · Five families of memory mechanisms
  3. The Transformer, the Engine Inside Every Modern ModelFoundational Modelling · The one big idea: attention
  4. Context Engineering and MemoryArchitecture and System Design · Real code: OpenClaw's memory search returns addresses, not files
  5. Reference Architecture and a Design ChecklistArchitecture and System Design · Context & prompts

Every term, A to Z