查询可见性如何改变 KV-Cache 压缩排名：匹配预算审核

arXiv:2607.11942v1 Announce Type: new Abstract: KV-cache compression methods are predominantly evaluated with the query appended to the context before compression -- a query-aware protocol.然而，压缩 KV 缓存的经济案例是重用：压缩文档一次，就可以回答许多未来针对它的问题。在该部署中，压缩必须与查询无关——在出现任何问题之前。 We present a matched-budget audit of six published compression methods against three trivial baselines on three open 7-9B models (144,300 paired evaluations on RULER-8192; 40,800 on LongBench; 50,000-resample paired bootstrap throughout). Everything is held fixed -- model, compression ratio, instances, decoding -- except the scoring rule.三项发现。 (1) Query visibility changes the rankings: under the agnostic protocol, of the five audited methods that share a common attention backend, only KeyDiff beats a best-of-3 trivial baseline consistently (31 of 36 cells), and the most widely deployed method, SnapKV, loses to "keep the start and the recent window" on average (-0.066). (2) The per-method drop between the two protocols is ordered consistently with how visible the question is to each method's scoring signal, legible in its source code: from Delta=+0.198 for SnapKV (the question sits inside its 64-token observation window) down to Delta=+0.011 for KeyDiff (its score contains no query term at all).