AI Paper Atlas

Research map / Architectures and efficiency / Quantization

PatternKV: Flattening KV Representation Expands Quantization Headroom

Ji Zhang, Yiwei Li, Shaoxiong Feng, Peiwen Yuan, Xinglin Wang, Yueqi Zhang, Jiayi Shi, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li

ICML 2026 · Architectures and efficiency › Quantization · approach: KV cache quantization · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
PatternKV pattern-aligned residual quantization instead of KIVIKIVI, Llama3.1-8B-Instruct · LongBench (2-bit)average score: 44.33 -> 45.33Improved
PatternKV instead of KIVI on 70BKIVI, Llama3.1-70B-Instruct · LongBench (2-bit)average score: 51.48 -> 51.61Improved
PatternKV instead of KIVI at 4-bitKIVI, Llama3.1-8B-Instruct · LongBench (4-bit)average score: 46.34 -> 46.41Improved
PatternKV instead of KIVI on long-CoT AIME 25KIVI, Llama-8B · AIME 25 (2-bit)Avg@8: 12.50 -> 17.50Improved
PatternKV instead of KIVI on Qwen-14B AIME 25KIVI, Qwen-14B · AIME 25 (2-bit)Avg@8: 37.08 -> 35.42Hurt
add K pattern alignment (ablation)PatternKV · LongBench (2-bit)average score: 44.53 -> 45.33Improved
add V pattern alignment (ablation)PatternKV · LongBench (2-bit)average score: 44.96 -> 45.33Improved
add new pattern vectors during decoding (ablation)PatternKV · LongBench (2-bit)average score: 45.37 -> 45.33No clear effect
add new pattern vectors during decoding (ablation)PatternKV · GSM8K (2-bit, zero-shot CoT)accuracy: 75.49 -> 75.58Improved
add adaptive z-test threshold for V patterns (ablation)PatternKV · LongBench (2-bit)average score: 24.67 -> 45.33Improved
add adaptive z-test threshold for V patterns (ablation)PatternKV · GSM8K (2-bit, zero-shot CoT)accuracy: 0.30 -> 75.58Improved
increase number of patterns from 2 to 32 (ablation)PatternKV · LongBench (2-bit)average score: 44.57 -> 45.33Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy