@inproceedings{maximov-etal-2026-2,
title = "From 2:4 to 8:16 sparsity patterns in {LLM}s for Outliers and Weights with Variance Correction",
author = "Maximov, Egor and
Kuzkina, Yulia and
Shvetsov, Egor and
Kanametov, Azamat and
Prutko, Aleksandr and
Zhelnin, Maxim and
Goncharov, Aleksei",
editor = "Li, Yunyao and
Rehm, Georg and
Tu, Mei",
booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics ({ACL} 2026)",
month = jul,
year = "2026",
address = "San Diego, California, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-industry.66/",
pages = "957--965",
ISBN = "979-8-89176-394-4",
abstract = "As large language models (LLMs) grow in size, efficient compression techniques like quantization and sparsification are critical. While quantization maintains performance with reduced precision, structured sparsity methods, such as N:M sparsification, often fall short due to limited flexibility and sensitivity to outlier weights. We explore 8:16 semi-structured sparsity, demonstrating its ability to surpass the Performance Threshold{---}where a compressed model matches the accuracy of its uncompressed or smaller counterpart under equivalent memory constraints. Compared to 2:4 sparsity, 8:16 offers greater flexibility with minimal storage overhead (0.875 vs. 0.75 bits/element). We also apply sparse structured patterns for salient weights, showing that structured sparsity for outliers is competitive with unstructured approaches, leading to equivalent or better results. Finally, we demonstrate that simple techniques such as variance correction and SmoothQuant-like weight equalization improve sparse models performance."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="maximov-etal-2026-2">
<titleInfo>
<title>From 2:4 to 8:16 sparsity patterns in LLMs for Outliers and Weights with Variance Correction</title>
</titleInfo>
<name type="personal">
<namePart type="given">Egor</namePart>
<namePart type="family">Maximov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yulia</namePart>
<namePart type="family">Kuzkina</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Egor</namePart>
<namePart type="family">Shvetsov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Azamat</namePart>
<namePart type="family">Kanametov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aleksandr</namePart>
<namePart type="family">Prutko</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Maxim</namePart>
<namePart type="family">Zhelnin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aleksei</namePart>
<namePart type="family">Goncharov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yunyao</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Georg</namePart>
<namePart type="family">Rehm</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mei</namePart>
<namePart type="family">Tu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-394-4</identifier>
</relatedItem>
<abstract>As large language models (LLMs) grow in size, efficient compression techniques like quantization and sparsification are critical. While quantization maintains performance with reduced precision, structured sparsity methods, such as N:M sparsification, often fall short due to limited flexibility and sensitivity to outlier weights. We explore 8:16 semi-structured sparsity, demonstrating its ability to surpass the Performance Threshold—where a compressed model matches the accuracy of its uncompressed or smaller counterpart under equivalent memory constraints. Compared to 2:4 sparsity, 8:16 offers greater flexibility with minimal storage overhead (0.875 vs. 0.75 bits/element). We also apply sparse structured patterns for salient weights, showing that structured sparsity for outliers is competitive with unstructured approaches, leading to equivalent or better results. Finally, we demonstrate that simple techniques such as variance correction and SmoothQuant-like weight equalization improve sparse models performance.</abstract>
<identifier type="citekey">maximov-etal-2026-2</identifier>
<location>
<url>https://aclanthology.org/2026.acl-industry.66/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>957</start>
<end>965</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T From 2:4 to 8:16 sparsity patterns in LLMs for Outliers and Weights with Variance Correction
%A Maximov, Egor
%A Kuzkina, Yulia
%A Shvetsov, Egor
%A Kanametov, Azamat
%A Prutko, Aleksandr
%A Zhelnin, Maxim
%A Goncharov, Aleksei
%Y Li, Yunyao
%Y Rehm, Georg
%Y Tu, Mei
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, USA
%@ 979-8-89176-394-4
%F maximov-etal-2026-2
%X As large language models (LLMs) grow in size, efficient compression techniques like quantization and sparsification are critical. While quantization maintains performance with reduced precision, structured sparsity methods, such as N:M sparsification, often fall short due to limited flexibility and sensitivity to outlier weights. We explore 8:16 semi-structured sparsity, demonstrating its ability to surpass the Performance Threshold—where a compressed model matches the accuracy of its uncompressed or smaller counterpart under equivalent memory constraints. Compared to 2:4 sparsity, 8:16 offers greater flexibility with minimal storage overhead (0.875 vs. 0.75 bits/element). We also apply sparse structured patterns for salient weights, showing that structured sparsity for outliers is competitive with unstructured approaches, leading to equivalent or better results. Finally, we demonstrate that simple techniques such as variance correction and SmoothQuant-like weight equalization improve sparse models performance.
%U https://aclanthology.org/2026.acl-industry.66/
%P 957-965
Markdown (Informal)
[From 2:4 to 8:16 sparsity patterns in LLMs for Outliers and Weights with Variance Correction](https://aclanthology.org/2026.acl-industry.66/) (Maximov et al., ACL 2026)
ACL
- Egor Maximov, Yulia Kuzkina, Egor Shvetsov, Azamat Kanametov, Aleksandr Prutko, Maxim Zhelnin, and Aleksei Goncharov. 2026. From 2:4 to 8:16 sparsity patterns in LLMs for Outliers and Weights with Variance Correction. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), pages 957–965, San Diego, California, USA. Association for Computational Linguistics.