How Binary Jumps and Local Connections Can Make Attention Faster
A popular-science introduction to sparse attention, binary information propagation, book and code experiments, and the gap between theoretical savings and GPU speed.
THE JOURNAL
Ideas, experiments, and observations in progress.
A popular-science introduction to sparse attention, binary information propagation, book and code experiments, and the gap between theoretical savings and GPU speed.
Why Opportunities Run Out—and What That Means for Work