What is an embedding?
Options
- A compressed copy of a document
- A hash that identifies a file
- The numeric id of a token
- A list of numbers (a vector) that represents the meaning of a text, so similar texts end up close together (Correct option)
Answer
A list of numbers (a vector) that represents the meaning of a text, so similar texts end up close together
Why
An embedding model turns a text into a vector of numbers. Texts with similar meaning get vectors that are close to each other, even if they use different words.
That makes it possible to search by meaning: embed the question, embed the documents, and look for the nearest vectors.
A hash is different: it changes completely when one letter changes and says nothing about meaning.
Remember
Embedding = a vector for meaning.