Open Models Almanac
Methodsmodel architecturetoken mixersparse attentionsparse attention indexer

implementation detail · filed under model architecture

Reuse QSA index selection across speculative decoding steps

Avoids repeating indexer work by reusing selected indices across speculative decoding steps.

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.

sources
2
model
1
lab adopt it
1
strongest
core

Documented in

Alibaba (Qwen)

Evidence

2 quotes from 2 sources, one card per model, each opening with its strongest.

How sources treat it
used 1core 1
↑