SOURCE-LINKED INTELLIGENCE

Jacap: Robust KV Cache Eviction via Jacobian-Based Nonlinear Information Capacity Preservation

arXiv · AI, language, vision and robotics · article · Sep 8, 2026 · UTC

Key-value (KV) cache eviction is essential for scaling long-context inference in Large Language Models. However, existing policies predominantly rely on empirical heuristics, lacking a rigorous characterization of token utility under the inherently nonlinear softmax attention mechanism. In this work, we rethink KV cache eviction through the lens of local information geometry, modeling the attention process as a nonlinear Gaussian communication channel. By performing a first-order Taylor expansion of the attention mapping, we derive the Jacobian Information Capacity, a novel objective that expl

Read original source ↗ Open in workspace

recordType

paper

region

Global