taxonomy node · level 4
fixed-pattern sparse attention
2 methods filed at this node or below it, from the sources of 2 models.
model architecture :: token mixer :: sparse attention :: fixed-pattern sparse attention
Matching aids for the classifier: local block; sparse global attention anchors; forced sink and local window.
In this branch 2
Everything filed at this node or below it, with one collapsible heading per child node.
filed here 2
By model
Which of this branch's techniques each model's own documents describe, and how strongly. Under each model: its strongest treatment anywhere in the branch.
| Model | techniques |
|---|---|
| NVIDIA-Nemotron-3-Ultra-550B-A55B core | Sparse Global Attention Anchors core— |
| MiniMax-M3 used | Local Block used— |