GAC: Noise-Aware Adaptive Mixing for Hybrid SFT-RL Post-Training

Yuelin Hu; Wei Liu; Zhenbo Yu; Zhengxue Cheng; Li Song (宋丽)

GAC: Noise-Aware Adaptive Mixing for Hybrid SFT-RL Post-Training

Yuelin Hu, Wei Liu, Zhenbo Yu, Zhengxue Cheng, Li Song

Correct Metadata for

Use this form to create a GitHub issue with structured data describing the correction. You will need a GitHub account. Once you create that issue, the correction will be reviewed by a staff member.

⚠️ Mobile Users: Submitting this form to create a new issue will only work with github.com, not the GitHub Mobile app.

Important: The Anthology treat PDFs as authoritative. Please use this form only to correct data that is out of line with the PDF. See our corrections guidelines if you need to change the PDF.

Title Adjust the title. Retain tags such as <fixed-case>.

Authors Adjust author names and order to match the PDF.

Abstract Correct abstract if needed. Retain XML formatting tags such as <tex-math>. You may use ... for bold, ... for italic, ... for underline, <sc>...</sc> for small-caps, <tt>...<tt> for typewriter text, <url>...</url> for URLs, <a href=...> for hyperlinks, and <par/> for paragraph breaks.

Verification against PDF Ensure that the new title/authors match the snapshot below. (If there is no snapshot or it is too small, consult the PDF.)

Authors concatenated from the text boxes above:

ALL author names match the snapshot above—including middle initials, hyphens, and accents.

Abstract

Hybrid post-training combining supervised fine-tuning (SFT) and reinforcement learning (RL) is the standard paradigm for aligning large language models, yet fixed mixing schedules cannot adapt when the relative noise of the two signals evolves. We derive a noise-aware mixing weight by minimizing an MSE upper bound on the mixed stochastic gradient, yielding a closed-form that balances gradient noise variance and SFT–RL disagreement. Building on the token-wise reweighting of CHORD (CITATION)

Anthology ID:: 2026.emnlp-main.119
Volume:: Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
Month:: October
Year:: 2026
Address:: Budapest, Hungary
Editors:: Hung-Yi Lee, Sunipa Dev, Jessy Li, Desmond Elliott
Venue:: EMNLP
SIG:: SIGDAT
Publisher:: Association for Computational Linguistics
Note:
Pages:: 2763–2779
Language:
URL:: https://aclanthology.org/2026.emnlp-main.119/
DOI:
Bibkey:
Cite (ACL):: Yuelin Hu, Wei Liu, Zhenbo Yu, Zhengxue Cheng, and Li Song. 2026. GAC: Noise-Aware Adaptive Mixing for Hybrid SFT-RL Post-Training. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 2763–2779, Budapest, Hungary. Association for Computational Linguistics.
Cite (Informal):: GAC: Noise-Aware Adaptive Mixing for Hybrid SFT-RL Post-Training (Hu et al., EMNLP 2026)
Copy Citation:
PDF:: https://aclanthology.org/2026.emnlp-main.119.pdf

PDF Cite Search Fix data

Export citation

BibTeX
MODS XML
Endnote
Preformatted

@inproceedings{hu-etal-2026-gac,
    title = "{GAC}: Noise-Aware Adaptive Mixing for Hybrid {SFT}-{RL} Post-Training",
    author = "Hu, Yuelin  and
      Liu, Wei  and
      Yu, Zhenbo  and
      Cheng, Zhengxue  and
      Song, Li",
    editor = "Lee, Hung-Yi  and
      Dev, Sunipa  and
      Li, Jessy  and
      Elliott, Desmond",
    booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
    month = oct,
    year = "2026",
    address = "Budapest, Hungary",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2026.emnlp-main.119/",
    pages = "2763--2779",
    ISBN = "979-8-89176-459-0",
    abstract = "Hybrid post-training combining supervised fine-tuning (SFT) and reinforcement learning (RL) is the standard paradigm for aligning large language models, yet fixed mixing schedules cannot adapt when the relative noise of the two signals evolves. We derive a noise-aware mixing weight by minimizing an MSE upper bound on the mixed stochastic gradient, yielding a closed-form that balances gradient noise variance and SFT{--}RL disagreement. Building on the token-wise reweighting of CHORD (CITATION)"
}

Download as File

<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="hu-etal-2026-gac">
    <titleInfo>
        <title>GAC: Noise-Aware Adaptive Mixing for Hybrid SFT-RL Post-Training</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Yuelin</namePart>
        <namePart type="family">Hu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Wei</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Zhenbo</namePart>
        <namePart type="family">Yu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Zhengxue</namePart>
        <namePart type="family">Cheng</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Li</namePart>
        <namePart type="family">Song</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>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</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">979-8-89176-459-0</identifier>
    </relatedItem>
    <abstract>Hybrid post-training combining supervised fine-tuning (SFT) and reinforcement learning (RL) is the standard paradigm for aligning large language models, yet fixed mixing schedules cannot adapt when the relative noise of the two signals evolves. We derive a noise-aware mixing weight by minimizing an MSE upper bound on the mixed stochastic gradient, yielding a closed-form that balances gradient noise variance and SFT–RL disagreement. Building on the token-wise reweighting of CHORD (CITATION)</abstract>
    <identifier type="citekey">hu-etal-2026-gac</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.119/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>2763</start>
            <end>2779</end>
        </extent>
    </part>
</mods>
</modsCollection>

Download as File

%0 Conference Proceedings
%T GAC: Noise-Aware Adaptive Mixing for Hybrid SFT-RL Post-Training
%A Hu, Yuelin
%A Liu, Wei
%A Yu, Zhenbo
%A Cheng, Zhengxue
%A Song, Li
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F hu-etal-2026-gac
%X Hybrid post-training combining supervised fine-tuning (SFT) and reinforcement learning (RL) is the standard paradigm for aligning large language models, yet fixed mixing schedules cannot adapt when the relative noise of the two signals evolves. We derive a noise-aware mixing weight by minimizing an MSE upper bound on the mixed stochastic gradient, yielding a closed-form that balances gradient noise variance and SFT–RL disagreement. Building on the token-wise reweighting of CHORD (CITATION)
%U https://aclanthology.org/2026.emnlp-main.119/
%P 2763-2779

Download as File

Markdown (Informal)

[GAC: Noise-Aware Adaptive Mixing for Hybrid SFT-RL Post-Training](https://aclanthology.org/2026.emnlp-main.119/) (Hu et al., EMNLP 2026)

GAC: Noise-Aware Adaptive Mixing for Hybrid SFT-RL Post-Training (Hu et al., EMNLP 2026)

ACL

Yuelin Hu, Wei Liu, Zhenbo Yu, Zhengxue Cheng, and Li Song. 2026. GAC: Noise-Aware Adaptive Mixing for Hybrid SFT-RL Post-Training. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 2763–2779, Budapest, Hungary. Association for Computational Linguistics.