paper-with-me

홈 › Papers

ManifoldKV: Training-Free KV Cache Compression via Euclidean Outlier Detection

2026-02-09 · Debajyoti Datta, Trishala Neeraj, Bibek Paudel, Vyom Sharma, Subhabrata Mukherjee arxiv

Long-context inference is constrained by KV-cache memory, which grows linearly with sequence length; KV-cache compression therefore hinges on reliably selecting which past tokens to retain. Most geometry-based eviction methods score keys by cosine similarity to a global centroid, but cosine is scale-invariant and can discard magnitude cues that distinguish semantically salient tokens. We propose ManifoldKV, a training-free scorer that ranks tokens by Euclidean distance to the key centroid, capturing both angular and radial deviations. On the RULER benchmark, ManifoldKV achieves 95.7% accuracy at 4K-16K contexts with 20% compression; matching the best geometric baseline while improving robustness in two regimes where cosine scoring fails. First, on multi-key retrieval, ManifoldKV reduces directional collisions, achieving 92.4% vs KeyDiff's 77.0% (+15.4 points) on 3-key NIAH at 50% compression. Second, to address dilution and performance collapse of global centroids at 64K context, we introduce WindowedManifoldKV, which restores accuracy to 84.3% at 25% compression, a 49-point recovery over global L2 and +3.2 points over KeyDiff. The method requires only 3 lines of code and works across 4 architectures without tuning.

📄 PDF Abstract BibTeX arXiv:2602.08343

Code (0)

등록된 구현이 없습니다.

Tasks

Outlier Detection

Similar Papers 제목 키워드 기반

VQKV: High-Fidelity and High-Ratio Cache Compression via Vector-Quantization

2026-03-17 · Yixuan Wang, Qingyu Shi, Jiayu Zhou, Dianbo Liu 외 arxiv

The growing context length of Large Language Models (LLMs) enlarges the Key-Value (KV) cache, limiting deployment in resource-limited environments. Prior training-free approaches for KV cache compression typically rely o…

R-KV: Redundancy-aware KV Cache Compression for Training-Free Reasoning Models Acceleration

2025-05-30 · Zefan Cai, Wen Xiao, Hanshi Sun, Cheng Luo 외

Reasoning models have demonstrated impressive performance in self-reflection and chain-of-thought reasoning. However, they often produce excessively long outputs, leading to prohibitively large key-value (KV) caches duri…

Mathematical Reasoning

ST-Lite: Training-Free KV Cache Compression with Spatio-Trajectory Guidance for Long-Horizon GUI Agents

2026-02-27 · Bowen Zhou, Zhou Xu, Wanli Li, Jingyu Xiao 외 arxiv

Training-free KV cache compression is essential for deploying vision-language GUI agents under memory and latency constraints, yet existing methods are designed for generic language workloads and ignore the distinctive s…

CSKV: Training-Efficient Channel Shrinking for KV Cache in Long-Context Scenarios

2024-09-16 · Luning Wang, Shiyao Li, Xuefei Ning, Zhihang Yuan 외

Large Language Models (LLMs) have been widely adopted to process long-context tasks. However, the large memory overhead of the key-value (KV) cache poses significant challenges in long-context scenarios. Existing trainin…

TaDA: Training-free recipe for Decoding with Adaptive KV Cache Compression and Mean-centering

2025-06-05 · Vinay Joshi, Pratik Prabhanjan Brahma, Zicheng Liu, Emad Barsoum

The key-value (KV) cache in transformer models is a critical component for efficient decoding or inference, yet its memory demands scale poorly with sequence length, posing a major challenge for scalable deployment of la…

Quantization