Focus learns a few centroids to gate long-range token attention, producing sparse attention that matches or beats full attention quality with up to 8.6x speedup at million-token lengths.
Flash A ttention-3: Fast and accurate attention with asynchrony and low-precision
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
CONDITIONAL 2representative citing papers
DualKV removes the N-fold recomputation of shared prompt tokens in RL policy updates via an exact two-region FlashAttention-style kernel, measuring 1.6–3.8x policy-update speedups.
citing papers explorer
-
Why Attend to Everything? Focus is the Key
Focus learns a few centroids to gate long-range token attention, producing sparse attention that matches or beats full attention quality with up to 8.6x speedup at million-token lengths.
-
DualKV: Shared-Prompt Flash Attention for Efficient RL Training with Large Rollouts and Long Contexts
DualKV removes the N-fold recomputation of shared prompt tokens in RL policy updates via an exact two-region FlashAttention-style kernel, measuring 1.6–3.8x policy-update speedups.