Preprint
Jul 2026
CoSA: Accelerating Long-Context Inference via Proxy-Kernel Co-Designed Sparse Attention
CoSA is proposed, a two-stage training-free Sparse Attention under proxy-kernel CO-design, which couples a Kernel-Aware Proxy (KAP) with an Ordered-Skipping Kernel (OSK) and achieves a 4.93% attention speedup and reduces end-to-end Time-to-First-Token by 2.53% with negligible performance degradation.
Yufei Xue, Lin Niu, Hong Liu et al.
· 0 citations