works on
efficiency 1large language models 1long-context inference 1proxy-kernel co-design 1sparse attention 1

From the 1 of 1 linked paper with an AI index.

collaborators
Showing cs.CLShow all

1 paper · 1 filter

cs.CL2026

CoSA: Accelerating Long-Context Inference via Proxy-Kernel Co-Designed Sparse Attention

Yufei Xue, Lin Niu, Hong Liu +6

CoSA introduces a training-free, two-stage sparse attention method that jointly designs a proxy and kernel to efficiently handle very long contexts, achieving faster inference with…