SOURCE-LINKED INTELLIGENCE

CEDAR: Error-Bounded Residual Routing for Efficient Long-Context Attention

arXiv · AI, language, vision and robotics · article · Sep 7, 2026 · UTC

Post-hoc sparse attention accelerates long-context prefill by routing each query to a small set of token-level interactions. Hard selection, however, assigns zero probability to every omitted chunk: a routing miss cannot be recovered, and a fixed expansion budget spends the same work on easy and ambiguous queries. We introduce Coarse-to-fine Error-aware Dynamic Attention Routing (CEDAR), a coarse-to-fine method that keeps the language model frozen while preserving global coverage. Each semantic chunk contributes a cheap key--value summary to a residual attention path; chunks with high estimate

Read original source ↗ Open in workspace

recordType

paper

region

Global