Quick answer
An embedding is a list of numbers that captures meaning. Similar ideas get similar numbers. This lets you search by meaning.
What it is
An embedding is a list of numbers. Each number captures a piece of meaning. Similar ideas get similar numbers.
Why it matters
Embeddings let you find related content without exact keyword matches. They power search, recommendations and RAG.
How to use it
- Pick an embedding model for your domain and language.
- Break content into small, self-contained chunks.
- Store each chunk as a vector in a database.
- At query time, embed the question and find the closest chunks.
Examples
- A support chatbot finds the most relevant help articles using embeddings.
- An e-commerce site recommends products by comparing embeddings of products and recent views.