/
LATENT REFERENCES / TAG2

self-attention

Archive curator and editor: Keigo Yoshida

This reference note belongs to Tag2 in Latent References, an archive curated by Keigo Yoshida. Its archive region is Learning models · Attention. The note preserves its source text and links so that readers can trace the material behind the 3D map.

Collection
Tag2
Archive region
Learning models · Attention

Archived reference note

English translation of the archived note. JP shows the original text. Source links and literal code are retained; the translation does not update or independently verify the source claims.

Self-Attention is something like a score of word-to-word relevance calculated using only words in a particular 1 sentence, a concept for associating words within a sentence. An input sentence is divided into words before being fed into the encoder. https://scrapbox.io/files/659a64cf15d31d0022565f96.png

Source updated 2024-01-07 · Snapshot 2026-10-08

Why this topic?

The text explains self-attention as relations between tokens.

This source-reviewed classification is an editorial interpretation. It does not replace calculated similarity or original source links.

Explore the same topic

These suggestions share an editorial topic. They are not a similarity score or a link asserted by the original note.

  • GrokSame source-reviewed topic · Learning models · Attention
  • WekinatorSame source-reviewed topic · Learning models · Attention
  • Gemini and Its AuthorsSame source-reviewed topic · Learning models · Attention