<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="mao-etal-2025-watermarking">
    <titleInfo>
        <title>Watermarking Large Language Models: An Unbiased and Low-risk Method</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Minjia</namePart>
        <namePart type="family">Mao</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Dongjun</namePart>
        <namePart type="family">Wei</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Zeyu</namePart>
        <namePart type="family">Chen</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Xiao</namePart>
        <namePart type="family">Fang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Michael</namePart>
        <namePart type="family">Chau</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 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Wanxiang</namePart>
            <namePart type="family">Che</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Joyce</namePart>
            <namePart type="family">Nabende</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ekaterina</namePart>
            <namePart type="family">Shutova</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Mohammad</namePart>
            <namePart type="given">Taher</namePart>
            <namePart type="family">Pilehvar</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">979-8-89176-251-0</identifier>
    </relatedItem>
    <abstract>Recent advancements in large language models (LLMs) have highlighted the risk of misusing them, raising the need for accurate detection of LLM-generated content. In response, a viable solution is to inject imperceptible identifiers into LLMs, known as watermarks. Our research extends the existing watermarking methods by proposing the novel Sampling One Then Accepting (STA-1) method. STA-1 is an unbiased watermark that preserves the original token distribution in expectation and has a lower risk of producing unsatisfactory outputs in low-entropy scenarios compared to existing unbiased watermarks. In watermark detection, STA-1 does not require prompts or a white-box LLM, provides statistical guarantees, demonstrates high efficiency in detection time, and remains robust against various watermarking attacks. Experimental results on low-entropy and high-entropy datasets demonstrate that STA-1 achieves the above properties simultaneously, making it a desirable solution for watermarking LLMs. Implementation codes for this study are available online.</abstract>
    <identifier type="citekey">mao-etal-2025-watermarking</identifier>
    <identifier type="doi">10.18653/v1/2025.acl-long.391</identifier>
    <location>
        <url>https://aclanthology.org/2025.acl-long.391/</url>
    </location>
    <part>
        <date>2025-07</date>
        <extent unit="page">
            <start>7939</start>
            <end>7960</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Watermarking Large Language Models: An Unbiased and Low-risk Method
%A Mao, Minjia
%A Wei, Dongjun
%A Chen, Zeyu
%A Fang, Xiao
%A Chau, Michael
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-251-0
%F mao-etal-2025-watermarking
%X Recent advancements in large language models (LLMs) have highlighted the risk of misusing them, raising the need for accurate detection of LLM-generated content. In response, a viable solution is to inject imperceptible identifiers into LLMs, known as watermarks. Our research extends the existing watermarking methods by proposing the novel Sampling One Then Accepting (STA-1) method. STA-1 is an unbiased watermark that preserves the original token distribution in expectation and has a lower risk of producing unsatisfactory outputs in low-entropy scenarios compared to existing unbiased watermarks. In watermark detection, STA-1 does not require prompts or a white-box LLM, provides statistical guarantees, demonstrates high efficiency in detection time, and remains robust against various watermarking attacks. Experimental results on low-entropy and high-entropy datasets demonstrate that STA-1 achieves the above properties simultaneously, making it a desirable solution for watermarking LLMs. Implementation codes for this study are available online.
%R 10.18653/v1/2025.acl-long.391
%U https://aclanthology.org/2025.acl-long.391/
%U https://doi.org/10.18653/v1/2025.acl-long.391
%P 7939-7960

Markdown (Informal)

[Watermarking Large Language Models: An Unbiased and Low-risk Method](https://aclanthology.org/2025.acl-long.391/) (Mao et al., ACL 2025)

• Watermarking Large Language Models: An Unbiased and Low-risk Method (Mao et al., ACL 2025)

ACL

• Minjia Mao, Dongjun Wei, Zeyu Chen, Xiao Fang, and Michael Chau. 2025. Watermarking Large Language Models: An Unbiased and Low-risk Method. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 7939–7960, Vienna, Austria. Association for Computational Linguistics.