AI Lab Tracker
Labs
Timeline
What's New
Collections
Sign in
Labs
Timeline
What's New
NSA: Native Sparse Attention
paper
2025-02-16
DeepSeek
Your tags
+ Collections
Create
Your notes
Add note
Hardware-aligned and natively trainable sparse attention mechanism.
Paper (arXiv)
Paper
Citations
2
arXiv
HTML
attention
architecture
efficiency
More Links
LinkedIn: Insights from Core Developer