Transformers · Level 3 of 5
Self-Attention
Attention in which queries, keys, and values come from the same sequence representation.
Each position can combine information from allowed positions in that sequence.
Example
A pronoun token gathers information from an earlier noun.
Listen to the definition and example
Audio transcript
Self-Attention. Attention in which queries, keys, and values come from the same sequence representation. Each position can combine information from allowed positions in that sequence. For example: A pronoun token gathers information from an earlier noun.
Explore this concept
Why it matters
This helps you trace how token representations exchange and transform information.
Start with
Related concepts
Quick recall question
Try answering before looking back at the definition.