Skip to content
AEO Canon · the reference for answer-engine optimization
AEO Glossary

Attention

Attention is the mechanism that lets a language model weigh which words in the input matter most for understanding each part, and it tends to concentrate on prominent, early, and clearly-related text.

BBurke Atkerson

Attention is how a model decides what to focus on. Within a transformer, the attention mechanism scores how relevant every word is to every other, so the model can emphasize the parts that carry meaning for the task — connecting a pronoun to its subject, a claim to its evidence, an answer to its question.

The practical AEO takeaway is that attention isn't evenly spread. Clear, prominent, well-structured statements draw more of it than buried or tangential text, which reinforces position bias and the case for answer-first, extractable writing. Make the sentence you want quoted prominent and unambiguous, and you make it the part the model attends to and reuses.

Example. In a paragraph that opens with a crisp answer and then elaborates, the model's attention lands on that opening claim — exactly the sentence you'd want an engine to lift and cite.

Relevant pillar

Related terms