@inproceedings{kulkarni-etal-2025-guir,
title = "{GUIR} at {S}em{E}val-2025 Task 4: Adaptive Weight Tuning with Gradual Negative Matching for {LLM} Unlearning",
author = "Kulkarni, Hrishikesh and
Goharian, Nazli and
Frieder, Ophir",
editor = "Rosenthal, Sara and
Ros{\'a}, Aiala and
Ghosh, Debanjan and
Zampieri, Marcos",
booktitle = "Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025)",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.semeval-1.152/",
pages = "1152--1158",
ISBN = "979-8-89176-273-2",
abstract = "Machine Unlearning for Large Language Models, referred to as LLM Unlearning is getting more and more attention as a result of regurgitation of sensitive and harmful content. In this paper, we present our method architecture, results, and analysis of our submission to Task4: Unlearning sensitive content from Large Language Models. This task includes three subtasks of LLM Unlearning on 1) Long Synthetic documents, 2) Short Synthetic documents, and 3) Real Training documents. Getting rid of the impact of undesirable and unauthorized responses is the core objective of unlearning. Furthermore, it is expected that unlearning should not have an adverse impact on the usability of the model. In this paper, we provide an approach for LLM unlearning that tries to make the model forget while maintaining usability of the model. We perform adaptive weight tuning with Gradient Ascent, KL minimization and Gradual Negative Matching loss functions. Our submission balances retain and forget abilities of the model while outperforming provided benchmarks."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kulkarni-etal-2025-guir">
<titleInfo>
<title>GUIR at SemEval-2025 Task 4: Adaptive Weight Tuning with Gradual Negative Matching for LLM Unlearning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hrishikesh</namePart>
<namePart type="family">Kulkarni</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nazli</namePart>
<namePart type="family">Goharian</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ophir</namePart>
<namePart type="family">Frieder</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sara</namePart>
<namePart type="family">Rosenthal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aiala</namePart>
<namePart type="family">Rosá</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Debanjan</namePart>
<namePart type="family">Ghosh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marcos</namePart>
<namePart type="family">Zampieri</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-273-2</identifier>
</relatedItem>
<abstract>Machine Unlearning for Large Language Models, referred to as LLM Unlearning is getting more and more attention as a result of regurgitation of sensitive and harmful content. In this paper, we present our method architecture, results, and analysis of our submission to Task4: Unlearning sensitive content from Large Language Models. This task includes three subtasks of LLM Unlearning on 1) Long Synthetic documents, 2) Short Synthetic documents, and 3) Real Training documents. Getting rid of the impact of undesirable and unauthorized responses is the core objective of unlearning. Furthermore, it is expected that unlearning should not have an adverse impact on the usability of the model. In this paper, we provide an approach for LLM unlearning that tries to make the model forget while maintaining usability of the model. We perform adaptive weight tuning with Gradient Ascent, KL minimization and Gradual Negative Matching loss functions. Our submission balances retain and forget abilities of the model while outperforming provided benchmarks.</abstract>
<identifier type="citekey">kulkarni-etal-2025-guir</identifier>
<location>
<url>https://aclanthology.org/2025.semeval-1.152/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>1152</start>
<end>1158</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T GUIR at SemEval-2025 Task 4: Adaptive Weight Tuning with Gradual Negative Matching for LLM Unlearning
%A Kulkarni, Hrishikesh
%A Goharian, Nazli
%A Frieder, Ophir
%Y Rosenthal, Sara
%Y Rosá, Aiala
%Y Ghosh, Debanjan
%Y Zampieri, Marcos
%S Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-273-2
%F kulkarni-etal-2025-guir
%X Machine Unlearning for Large Language Models, referred to as LLM Unlearning is getting more and more attention as a result of regurgitation of sensitive and harmful content. In this paper, we present our method architecture, results, and analysis of our submission to Task4: Unlearning sensitive content from Large Language Models. This task includes three subtasks of LLM Unlearning on 1) Long Synthetic documents, 2) Short Synthetic documents, and 3) Real Training documents. Getting rid of the impact of undesirable and unauthorized responses is the core objective of unlearning. Furthermore, it is expected that unlearning should not have an adverse impact on the usability of the model. In this paper, we provide an approach for LLM unlearning that tries to make the model forget while maintaining usability of the model. We perform adaptive weight tuning with Gradient Ascent, KL minimization and Gradual Negative Matching loss functions. Our submission balances retain and forget abilities of the model while outperforming provided benchmarks.
%U https://aclanthology.org/2025.semeval-1.152/
%P 1152-1158
Markdown (Informal)
[GUIR at SemEval-2025 Task 4: Adaptive Weight Tuning with Gradual Negative Matching for LLM Unlearning](https://aclanthology.org/2025.semeval-1.152/) (Kulkarni et al., SemEval 2025)
ACL