Open Models Almanac
Methodsmodel architecturetoken mixersparse attentionsparse attention indexer

single mention · model architecture

Shared DSA indexer across sliding-window layers

Shares a single DSA indexer among the former sliding-window layers within each attention group.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See sparse attention indexer for the methods filed beside it.

Filed under model architecture › token mixer › sparse attention › sparse attention indexer