Model techniques map
Techniquesmodel architecturetoken mixersparse attentionblock selection

specific method · filed under model architecture

Mixture of Block Attention

A sparse-attention method named for a mixture of block attention; the supplied evidence gives no further mechanism details.

Also called MoBA.

source
1
model
1
labs adopt it
0
strongest
evaluated

How sources treat it

One count per evidence span, weakest treatment to strongest.

evaluated 1

Documented in

Evidence

1 span quoted from the sources, strongest treatment first.

Compared with approaches like DSA and MoBA, MSA can partition the KV into blocks more precisely

evaluatedevaluation onlyin MiniMax-M3MiniMax

Filed alongside

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