specific method · filed under model architecture
Sparse Global Attention Anchors
Sparse global attention anchors used in a hybrid Mamba-2 stack, distinct from the always-selected local block.
- 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.
a hybrid Mamba-2 stack with sparse global Attention anchors
coremodel architecturein Nemotron 3 UltraNVIDIA
Filed alongside
Other methods under model architecture :: token mixer :: sparse attention :: fixed-pattern sparse attention.