Model techniques map
Techniquesinference & servingKV cache management

implementation detail · filed under inference & serving

Mamba prefix caching in align mode

Uses prefix caching for Mamba in experimental align mode.

Also called prefix caching in align mode for Mamba.

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.

Prefix Caching (Mamba): currently experimental in "align" mode.

evaluatedunclearin Qwen3.6-35B-A3BQwen

Filed alongside

Other methods under inference & serving :: KV cache management.