@inproceedings{moon-etal-2025-limacost,
title = "{L}ima{C}ost: Data Valuation for Instruction Tuning of Large Language Models",
author = "Moon, Hyeonseok and
Seo, Jaehyung and
Koo, Seonmin and
Kim, Jinsung and
Ham, Young-kyoung and
Moon, Jiwon and
Lim, Heuiseok",
editor = "Christodoulopoulos, Christos and
Chakraborty, Tanmoy and
Rose, Carolyn and
Peng, Violet",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2025",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-emnlp.688/",
doi = "10.18653/v1/2025.findings-emnlp.688",
pages = "12841--12854",
ISBN = "979-8-89176-335-7",
abstract = "Instruction tuning (IT) is an effective approach for aligning large language models (LLMs) with human intentions. There is ongoing discourse regarding the data quality for IT. As an effort to find the robust criteria of data quality for IT, we introduce LimaCost, a data quality measure that exhibits a strong correlation with model performance. LimaCost utilizes LIMA dataset, which effectiveness in IT has already been validated by several previous works. LimaCost then estimates the value of a given data by estimating how many LIMA data points might be needed to approximate its gradient. Our experiments reveal that LimaCost enables effective data selection that derive high alignment performance. We demonstrate that selecting data based on high LimaCost proves to be more effective than existing data selection strategies."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="moon-etal-2025-limacost">
<titleInfo>
<title>LimaCost: Data Valuation for Instruction Tuning of Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hyeonseok</namePart>
<namePart type="family">Moon</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jaehyung</namePart>
<namePart type="family">Seo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Seonmin</namePart>
<namePart type="family">Koo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jinsung</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Young-kyoung</namePart>
<namePart type="family">Ham</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiwon</namePart>
<namePart type="family">Moon</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Heuiseok</namePart>
<namePart type="family">Lim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Christos</namePart>
<namePart type="family">Christodoulopoulos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tanmoy</namePart>
<namePart type="family">Chakraborty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Carolyn</namePart>
<namePart type="family">Rose</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Violet</namePart>
<namePart type="family">Peng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Suzhou, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-335-7</identifier>
</relatedItem>
<abstract>Instruction tuning (IT) is an effective approach for aligning large language models (LLMs) with human intentions. There is ongoing discourse regarding the data quality for IT. As an effort to find the robust criteria of data quality for IT, we introduce LimaCost, a data quality measure that exhibits a strong correlation with model performance. LimaCost utilizes LIMA dataset, which effectiveness in IT has already been validated by several previous works. LimaCost then estimates the value of a given data by estimating how many LIMA data points might be needed to approximate its gradient. Our experiments reveal that LimaCost enables effective data selection that derive high alignment performance. We demonstrate that selecting data based on high LimaCost proves to be more effective than existing data selection strategies.</abstract>
<identifier type="citekey">moon-etal-2025-limacost</identifier>
<identifier type="doi">10.18653/v1/2025.findings-emnlp.688</identifier>
<location>
<url>https://aclanthology.org/2025.findings-emnlp.688/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>12841</start>
<end>12854</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T LimaCost: Data Valuation for Instruction Tuning of Large Language Models
%A Moon, Hyeonseok
%A Seo, Jaehyung
%A Koo, Seonmin
%A Kim, Jinsung
%A Ham, Young-kyoung
%A Moon, Jiwon
%A Lim, Heuiseok
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Findings of the Association for Computational Linguistics: EMNLP 2025
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-335-7
%F moon-etal-2025-limacost
%X Instruction tuning (IT) is an effective approach for aligning large language models (LLMs) with human intentions. There is ongoing discourse regarding the data quality for IT. As an effort to find the robust criteria of data quality for IT, we introduce LimaCost, a data quality measure that exhibits a strong correlation with model performance. LimaCost utilizes LIMA dataset, which effectiveness in IT has already been validated by several previous works. LimaCost then estimates the value of a given data by estimating how many LIMA data points might be needed to approximate its gradient. Our experiments reveal that LimaCost enables effective data selection that derive high alignment performance. We demonstrate that selecting data based on high LimaCost proves to be more effective than existing data selection strategies.
%R 10.18653/v1/2025.findings-emnlp.688
%U https://aclanthology.org/2025.findings-emnlp.688/
%U https://doi.org/10.18653/v1/2025.findings-emnlp.688
%P 12841-12854
Markdown (Informal)
[LimaCost: Data Valuation for Instruction Tuning of Large Language Models](https://aclanthology.org/2025.findings-emnlp.688/) (Moon et al., Findings 2025)
ACL