@inproceedings{li-etal-2026-kurtosis,
title = "Kurtosis-Aware Coupled Sparsity: Training-Free Activation Sparsity for Large Language Models",
author = "Li, Guangyan and
Sun, Luoyang and
Li, Siyuan and
Zhang, Haifeng and
Tang, Yongqiang and
Zhang, Wensheng",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1884/",
pages = "42942--42961",
ISBN = "979-8-89176-459-0",
abstract = "The massive scale of Large Language Models (LLMs) incurs prohibitive computational costs, limiting their ubiquitous deployment. While activation sparsity is a promising training-free solution, existing magnitude-based methods overlook weight structures and dependencies. We reveal that standard weight norms fail as importance proxies, identifying instead that weight distribution heavy-tailedness is critical. To this end, we propose Kurtosis-Aware Coupled Sparsity (KACS), a novel training-free framework that introduces a Kurtosis-based metric to explicitly capture and prioritize outlier-rich weight columns. Furthermore, we develop an Interaction-Aware evaluation mechanism that assesses the joint importance of structurally coupled projections (e.g., Gate-Up and Q-K-V), ensuring information retention across interacting pathways. In addition, an adaptive sparsity allocation strategy is employed to distribute the sparsity budget across layers and components. Extensive experiments on Llama and Mistral models demonstrate that KACS consistently outperforms state-of-the-art baselines, retaining over 97{\%} of the original performance at 50{\%} sparsity."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-etal-2026-kurtosis">
<titleInfo>
<title>Kurtosis-Aware Coupled Sparsity: Training-Free Activation Sparsity for Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Guangyan</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Luoyang</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Siyuan</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haifeng</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yongqiang</namePart>
<namePart type="family">Tang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wensheng</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>The massive scale of Large Language Models (LLMs) incurs prohibitive computational costs, limiting their ubiquitous deployment. While activation sparsity is a promising training-free solution, existing magnitude-based methods overlook weight structures and dependencies. We reveal that standard weight norms fail as importance proxies, identifying instead that weight distribution heavy-tailedness is critical. To this end, we propose Kurtosis-Aware Coupled Sparsity (KACS), a novel training-free framework that introduces a Kurtosis-based metric to explicitly capture and prioritize outlier-rich weight columns. Furthermore, we develop an Interaction-Aware evaluation mechanism that assesses the joint importance of structurally coupled projections (e.g., Gate-Up and Q-K-V), ensuring information retention across interacting pathways. In addition, an adaptive sparsity allocation strategy is employed to distribute the sparsity budget across layers and components. Extensive experiments on Llama and Mistral models demonstrate that KACS consistently outperforms state-of-the-art baselines, retaining over 97% of the original performance at 50% sparsity.</abstract>
<identifier type="citekey">li-etal-2026-kurtosis</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1884/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>42942</start>
<end>42961</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Kurtosis-Aware Coupled Sparsity: Training-Free Activation Sparsity for Large Language Models
%A Li, Guangyan
%A Sun, Luoyang
%A Li, Siyuan
%A Zhang, Haifeng
%A Tang, Yongqiang
%A Zhang, Wensheng
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F li-etal-2026-kurtosis
%X The massive scale of Large Language Models (LLMs) incurs prohibitive computational costs, limiting their ubiquitous deployment. While activation sparsity is a promising training-free solution, existing magnitude-based methods overlook weight structures and dependencies. We reveal that standard weight norms fail as importance proxies, identifying instead that weight distribution heavy-tailedness is critical. To this end, we propose Kurtosis-Aware Coupled Sparsity (KACS), a novel training-free framework that introduces a Kurtosis-based metric to explicitly capture and prioritize outlier-rich weight columns. Furthermore, we develop an Interaction-Aware evaluation mechanism that assesses the joint importance of structurally coupled projections (e.g., Gate-Up and Q-K-V), ensuring information retention across interacting pathways. In addition, an adaptive sparsity allocation strategy is employed to distribute the sparsity budget across layers and components. Extensive experiments on Llama and Mistral models demonstrate that KACS consistently outperforms state-of-the-art baselines, retaining over 97% of the original performance at 50% sparsity.
%U https://aclanthology.org/2026.emnlp-main.1884/
%P 42942-42961
Markdown (Informal)
[Kurtosis-Aware Coupled Sparsity: Training-Free Activation Sparsity for Large Language Models](https://aclanthology.org/2026.emnlp-main.1884/) (Li et al., EMNLP 2026)
ACL