全局计算，局部物化：稀疏事件KV的内存契约

AI 摘要

该论文研究了长时程智能体在KV缓存重用中，被保留的缓存条目在源事件被移除后是否仍具有信息量。通过省略早期观察的对照实验，发现下游事件的缓存行能独立反映已消失输入的计算结果，称为语义物化。论文还展示了通过精心措辞的载体事件可将捐赠者对齐恢复率从6%提升至51%，并总结了稀疏事件KV服务的记忆契约。

正文节选

Compute Globally, Materialize Locally: The Memory Contract of Sparse Event-KV Abstract Long-horizon agents increasingly reuse their KV cache as memory: a serving system keeps a subset of cached entries and drops the rest. Eviction and episodic-memory schemes therefore rest on a premise rarely tested directly, that a retained event is still informative once the observations that produced it are gone. We test it by omitting one earlier observation from what is served, across otherwise identical ag

发布时间：—

抓取时间：2026-08-04 23:51

来源机构：Hugging Face