Model techniques map
Techniquesmodel architecturetoken mixersparse attention

specific method · filed under model architecture

Cross-layer KV and index reuse with statically assigned CSA2 modes

source
1
model
1
lab adopt it
1
strongest
core

How sources treat it

One count per evidence span, weakest treatment to strongest.

core 1

Documented in

Evidence

1 span quoted from the sources, strongest treatment first.

CSA2 has three statically assigned modes: Full, Reindex, and Reuse. Reuse Mode reuses both global KV and the Top-K indices in a preceding layer, and directly performs sparse attention.

coremodel architecturein DeepSeek-V4.1-FlashDeepSeek

Filed alongside

Other methods under model architecture :: token mixer :: sparse attention.