Model techniques map
Techniquesmodel architecturetoken mixersparse attention

implementation detail · filed under model architecture

Sequential block processing

An implementation detail in which sparse-attention blocks are processed sequentially.

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.

MSA flips the logic and processes blocks sequentially.

coresoftware implementationin MiniMax-M3MiniMax

Filed alongside

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