Last released Feb 6, 2026
Efficient sparse attention kernels for long-context language models
Supported by