@inproceedings{makhov-etal-2026-cospadi,
title = "{C}o{S}pa{D}i: Compressing {LLM}s via Calibration-Guided Sparse Dictionary Learning",
author = "Makhov, Denis and
Shopkhoev, Dmitriy and
Zhussip, Magauiya and
Ali, Ammar and
Lefkimmiatis, Stamatios",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.284/",
pages = "6005--6031",
ISBN = "979-8-89176-462-0",
abstract = "Post-training LLM compression often relies on low-rank approximations, which force all columns of a projection matrix to share a single low-dimensional subspace. We propose *CoSpaDi*, a training-free compression framework that replaces this single-subspace assumption with a union-of-subspaces model via sparse dictionary learning. *CoSpaDi* factorizes each weight matrix into a dense dictionary and column-sparse coefficients, allowing different columns to select different subsets of dictionary atoms at the same storage budget. To preserve model behavior, we use calibration activations to transform functional reconstruction into a standard dictionary learning problem. Across Llama and Qwen models, CoSpaDi improves accuracy-compression and perplexity-compression trade-offs over SVD-based and structured pruning baselines at 20-40{\%} compression ratios, while naturally supporting sparse-dense computation and post-training quantization of sparse coefficients."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="makhov-etal-2026-cospadi">
<titleInfo>
<title>CoSpaDi: Compressing LLMs via Calibration-Guided Sparse Dictionary Learning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Denis</namePart>
<namePart type="family">Makhov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dmitriy</namePart>
<namePart type="family">Shopkhoev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Magauiya</namePart>
<namePart type="family">Zhussip</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ammar</namePart>
<namePart type="family">Ali</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Stamatios</namePart>
<namePart type="family">Lefkimmiatis</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Post-training LLM compression often relies on low-rank approximations, which force all columns of a projection matrix to share a single low-dimensional subspace. We propose *CoSpaDi*, a training-free compression framework that replaces this single-subspace assumption with a union-of-subspaces model via sparse dictionary learning. *CoSpaDi* factorizes each weight matrix into a dense dictionary and column-sparse coefficients, allowing different columns to select different subsets of dictionary atoms at the same storage budget. To preserve model behavior, we use calibration activations to transform functional reconstruction into a standard dictionary learning problem. Across Llama and Qwen models, CoSpaDi improves accuracy-compression and perplexity-compression trade-offs over SVD-based and structured pruning baselines at 20-40% compression ratios, while naturally supporting sparse-dense computation and post-training quantization of sparse coefficients.</abstract>
<identifier type="citekey">makhov-etal-2026-cospadi</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.284/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>6005</start>
<end>6031</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T CoSpaDi: Compressing LLMs via Calibration-Guided Sparse Dictionary Learning
%A Makhov, Denis
%A Shopkhoev, Dmitriy
%A Zhussip, Magauiya
%A Ali, Ammar
%A Lefkimmiatis, Stamatios
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F makhov-etal-2026-cospadi
%X Post-training LLM compression often relies on low-rank approximations, which force all columns of a projection matrix to share a single low-dimensional subspace. We propose *CoSpaDi*, a training-free compression framework that replaces this single-subspace assumption with a union-of-subspaces model via sparse dictionary learning. *CoSpaDi* factorizes each weight matrix into a dense dictionary and column-sparse coefficients, allowing different columns to select different subsets of dictionary atoms at the same storage budget. To preserve model behavior, we use calibration activations to transform functional reconstruction into a standard dictionary learning problem. Across Llama and Qwen models, CoSpaDi improves accuracy-compression and perplexity-compression trade-offs over SVD-based and structured pruning baselines at 20-40% compression ratios, while naturally supporting sparse-dense computation and post-training quantization of sparse coefficients.
%U https://aclanthology.org/2026.findings-emnlp.284/
%P 6005-6031
Markdown (Informal)
[CoSpaDi: Compressing LLMs via Calibration-Guided Sparse Dictionary Learning](https://aclanthology.org/2026.findings-emnlp.284/) (Makhov et al., Findings 2026)
ACL