Periodic Weak Spots: Phase Sensitivity from Chunked KV-Cache Compression

Chunked KV-cache compression reduces the memory and attention costs of long-context inference by compressing windows of consecutive tokens into fewer cache entries at a fixed stride.

By Zhu, Pu, Yi +6arXivSep 28, 2026

Score█████░░░░░5.5

VerdictCompetent work. Briefs at most.

Read the original PDF

Abstract

Chunked KV-cache compression reduces the memory and attention costs of long-context inference by compressing windows of consecutive tokens into fewer cache entries at a fixed stride. Such compression also introduces a new positional coordinate: a token's phase, or its position relative to compression-window boundaries. We uncover a systematic asymmetry in models using such compression: the same information can be easy to retrieve at one phase and difficult at another. We call this periodic variation in retrieval performance phase sensitivity. In large open-weight models with such compression, long-context retrieval accuracy can differ by up to 40 percentage points across phases, revealing periodic weak spots that average benchmark scores can conceal. To investigate this behavior, we pretrain a family of transformers from scratch across multiple KV-compression designs, reproducing phase sensitivity across the variants. Mechanistic analysis using causal interventions in these models reveals phase specialization: different attention components contribute asymmetrically to retrieving information at different source phases. We further analyze idealized retrieval models, showing how gradient flow dynamics may favor sharp phase specialization. Evaluating models with chunked KV-cache compression thus requires measuring across compression phases: high average accuracy can coexist with systematic positional failures.

Xingyu Zhu, Pu, Yi, Ziheng Cheng, Ang Lv, Jing Liu, Lexing Ying, Yiyuan Ma, Xin Dong

The editor's rubric

Heuristic review

Dimension	Level	Weight	What that level means
Leverage	███░░ 3	24%	A method or resource many groups across the field will adopt within a year.
Magnitude	██░░░ 2	18%	Solid incremental gain on a meaningful problem.
Evidence	███░░ 3	14%	Solid: multiple benchmarks or cohorts, ablations, fair baselines, released code or data.
Novelty	███░░ 3	16%	A genuinely new approach to an open problem.
Trajectory	██░░░ 2	18%	Some room to improve with obvious engineering.
Stakes	██░░░ 2	10%	Benefits a professional community (practitioners, clinicians, engineers).

Editor’s rationale

Heuristic triage from title and abstract text only, not a reading of the paper. Cues found: method (named contribution); novelty (discovery); verification (error bars).

How the score was computed

rank-2026-09-29

Score█████░░░░░5.5

Score = 10 × (65% × adjusted merit / 10 + 25% × attention + 10% × freshness)

Merit: 5.1 / 10; Weighted rubric, evidence-gated.
Adjusted merit: 4.4 / 10; Shrunk toward the desk prior by editor confidence (36%).
Attention: 87%; Citations, upvotes, points, mentions.
Freshness: 42%; Half-life decay since publication.

Citations0 (reference 15, via semantic-scholar, Oct 1, 2026, 02:16 UTC)
Influential citations0 (reference 3, via semantic-scholar, Oct 1, 2026, 02:16 UTC)
Hugging Face upvotes86 (reference 25, via hf-daily, Oct 1, 2026, 02:16 UTC)

The record

Reviewed by heuristic-v2 on Oct 1, 2026, 02:16 UTC. Paper type: method.
BRIEF, No.10 in the Artificial Intelligence edition of October 1, 2026.