arXiv:2611.12956v1 Announce Type: new Abstract: DeepSeek Sparse Attention (DSA) reduces the computational cost of attention through token-level selection. However, its indexer still scores the entire causal prefix for each query, making indexing a dominant source of latency in long-context inference. Notably, the indexer alone can account for more than 50% of total prefill GPU time at long context lengths. To address this bottleneck, we propose ADIndex, an Adaptive and Distillation-driven Indexing method for token-level sparse attention.
Machine Learning · 2026-11-11
ADIndex: Layer-Adaptive Sparse Attention Indexing via Group-Level Distillation
arXiv:2611.12956v1 Announce Type: new Abstract: DeepSeek Sparse Attention (DSA) reduces the computational cost of attention through token-level selection. However, its indexer still scores the entire causal prefix for each query, making indexing a dominant source of latency in long-context inference. Notably, the indexer alone can account for more than 50% of total prefill GPU time at long context lengths.
Source and attribution
This is an attributed editorial digest. For the complete study and original reporting, follow the source link.
Reader feedback
0 likes · 0 dislikes