@inproceedings{cheng-etal-2025-weaponization,
title = "On Weaponization-Resistant Large Language Models with Prospect Theoretic Alignment",
author = "Cheng, Zehua and
Zhang, Manying and
Sun, Jiahao and
Dai, Wei",
editor = "Rambow, Owen and
Wanner, Leo and
Apidianaki, Marianna and
Al-Khalifa, Hend and
Eugenio, Barbara Di and
Schockaert, Steven",
booktitle = "Proceedings of the 31st International Conference on Computational Linguistics",
month = jan,
year = "2025",
address = "Abu Dhabi, UAE",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.coling-main.687/",
pages = "10309--10324",
abstract = "Large language models (LLMs) have made significant advancements, but their increasing capabilities present serious risks of misuse, particularly in open-weight models where direct access to the model`s parameters is possible. Current safeguards, designed for closed-weight API models, are inadequate for open-weight models, as minimal fine-tuning can bypass these protections. Preserving the integrity of open-weight LLMs before deployment has thus become a critical challenge. We argue that these vulnerabilities stem from the overemphasis on maximizing the LLM`s log-likelihood during training, which amplifies data biases, especially with large datasets. To address these issues, we introduce Kahneman and Tversky`s Prospect Theoretic Integrity Preserving Alignment (KT-IPA), a framework that prioritizes maximizing generative utility rather than a singular optimization metric. This approach strengthens LLMs against misuse and weaponization while maintaining high performance, even after extensive fine-tuning. Our results demonstrate that integrating prospect theory into LLM training enhances robustness, security, and responsible innovation in this rapidly evolving field. Our codes are available on https://anonymous.4open.science/r/KT-IPA-40B7"
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="cheng-etal-2025-weaponization">
<titleInfo>
<title>On Weaponization-Resistant Large Language Models with Prospect Theoretic Alignment</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zehua</namePart>
<namePart type="family">Cheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Manying</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiahao</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wei</namePart>
<namePart type="family">Dai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-01</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 31st International Conference on Computational Linguistics</title>
</titleInfo>
<name type="personal">
<namePart type="given">Owen</namePart>
<namePart type="family">Rambow</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leo</namePart>
<namePart type="family">Wanner</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marianna</namePart>
<namePart type="family">Apidianaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hend</namePart>
<namePart type="family">Al-Khalifa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Barbara</namePart>
<namePart type="given">Di</namePart>
<namePart type="family">Eugenio</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Steven</namePart>
<namePart type="family">Schockaert</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Abu Dhabi, UAE</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Large language models (LLMs) have made significant advancements, but their increasing capabilities present serious risks of misuse, particularly in open-weight models where direct access to the model‘s parameters is possible. Current safeguards, designed for closed-weight API models, are inadequate for open-weight models, as minimal fine-tuning can bypass these protections. Preserving the integrity of open-weight LLMs before deployment has thus become a critical challenge. We argue that these vulnerabilities stem from the overemphasis on maximizing the LLM‘s log-likelihood during training, which amplifies data biases, especially with large datasets. To address these issues, we introduce Kahneman and Tversky‘s Prospect Theoretic Integrity Preserving Alignment (KT-IPA), a framework that prioritizes maximizing generative utility rather than a singular optimization metric. This approach strengthens LLMs against misuse and weaponization while maintaining high performance, even after extensive fine-tuning. Our results demonstrate that integrating prospect theory into LLM training enhances robustness, security, and responsible innovation in this rapidly evolving field. Our codes are available on https://anonymous.4open.science/r/KT-IPA-40B7</abstract>
<identifier type="citekey">cheng-etal-2025-weaponization</identifier>
<location>
<url>https://aclanthology.org/2025.coling-main.687/</url>
</location>
<part>
<date>2025-01</date>
<extent unit="page">
<start>10309</start>
<end>10324</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T On Weaponization-Resistant Large Language Models with Prospect Theoretic Alignment
%A Cheng, Zehua
%A Zhang, Manying
%A Sun, Jiahao
%A Dai, Wei
%Y Rambow, Owen
%Y Wanner, Leo
%Y Apidianaki, Marianna
%Y Al-Khalifa, Hend
%Y Eugenio, Barbara Di
%Y Schockaert, Steven
%S Proceedings of the 31st International Conference on Computational Linguistics
%D 2025
%8 January
%I Association for Computational Linguistics
%C Abu Dhabi, UAE
%F cheng-etal-2025-weaponization
%X Large language models (LLMs) have made significant advancements, but their increasing capabilities present serious risks of misuse, particularly in open-weight models where direct access to the model‘s parameters is possible. Current safeguards, designed for closed-weight API models, are inadequate for open-weight models, as minimal fine-tuning can bypass these protections. Preserving the integrity of open-weight LLMs before deployment has thus become a critical challenge. We argue that these vulnerabilities stem from the overemphasis on maximizing the LLM‘s log-likelihood during training, which amplifies data biases, especially with large datasets. To address these issues, we introduce Kahneman and Tversky‘s Prospect Theoretic Integrity Preserving Alignment (KT-IPA), a framework that prioritizes maximizing generative utility rather than a singular optimization metric. This approach strengthens LLMs against misuse and weaponization while maintaining high performance, even after extensive fine-tuning. Our results demonstrate that integrating prospect theory into LLM training enhances robustness, security, and responsible innovation in this rapidly evolving field. Our codes are available on https://anonymous.4open.science/r/KT-IPA-40B7
%U https://aclanthology.org/2025.coling-main.687/
%P 10309-10324
Markdown (Informal)
[On Weaponization-Resistant Large Language Models with Prospect Theoretic Alignment](https://aclanthology.org/2025.coling-main.687/) (Cheng et al., COLING 2025)
ACL