IndexCache: Accelerating Sparse Attention via Cross-Layer Index Reuse Paper • 2603.12201 • Published 2 days ago • 35
SpargeAttention2: Trainable Sparse Attention via Hybrid Top-k+Top-p Masking and Distillation Fine-Tuning Paper • 2602.13515 • Published 29 days ago • 43
DZ-TDPO: Non-Destructive Temporal Alignment for Mutable State Tracking in Long-Context Dialogue Paper • 2512.03704 • Published Dec 3, 2025 • 2