<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="hu-etal-2026-navigating">
    <titleInfo>
        <title>Navigating the Alignment-Calibration Trade-off: A Pareto-Superior Frontier via Model Merging</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Tiancheng</namePart>
        <namePart type="family">Hu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Benjamin</namePart>
        <namePart type="family">Minixhofer</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Nigel</namePart>
        <namePart type="family">Collier</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-07</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: ACL 2026</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Maria</namePart>
            <namePart type="family">Liakata</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Viviane</namePart>
            <namePart type="given">P</namePart>
            <namePart type="family">Moreira</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jiajun</namePart>
            <namePart type="family">Zhang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">David</namePart>
            <namePart type="family">Jurgens</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">San Diego, California, United States</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-395-1</identifier>
    </relatedItem>
    <abstract>The “alignment tax” of post-training is typically framed as a drop in task accuracy. We show it also involves a severe loss of calibration, making models overconfident, less reliable, and model outputs less diverse. We demonstrate that this trade-off can be navigated effectively via a simple post-hoc intervention: interpolating between a model’s weights before and after alignment. Crucially, this is not a strict trade-off. We find that the process consistently reveals Pareto-optimal interpolations—models that improve accuracy beyond both parents while substantially recovering the calibration lost during alignment. Our work demonstrates that simple model merging provides a computationally efficient method for mitigating the full scope of the alignment tax, yielding models that are more capable and more reliable.</abstract>
    <identifier type="citekey">hu-etal-2026-navigating</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-acl.2104/</url>
    </location>
    <part>
        <date>2026-07</date>
        <extent unit="page">
            <start>42405</start>
            <end>42422</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Navigating the Alignment-Calibration Trade-off: A Pareto-Superior Frontier via Model Merging
%A Hu, Tiancheng
%A Minixhofer, Benjamin
%A Collier, Nigel
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F hu-etal-2026-navigating
%X The “alignment tax” of post-training is typically framed as a drop in task accuracy. We show it also involves a severe loss of calibration, making models overconfident, less reliable, and model outputs less diverse. We demonstrate that this trade-off can be navigated effectively via a simple post-hoc intervention: interpolating between a model’s weights before and after alignment. Crucially, this is not a strict trade-off. We find that the process consistently reveals Pareto-optimal interpolations—models that improve accuracy beyond both parents while substantially recovering the calibration lost during alignment. Our work demonstrates that simple model merging provides a computationally efficient method for mitigating the full scope of the alignment tax, yielding models that are more capable and more reliable.
%U https://aclanthology.org/2026.findings-acl.2104/
%P 42405-42422

Markdown (Informal)

[Navigating the Alignment-Calibration Trade-off: A Pareto-Superior Frontier via Model Merging](https://aclanthology.org/2026.findings-acl.2104/) (Hu et al., Findings 2026)

• Navigating the Alignment-Calibration Trade-off: A Pareto-Superior Frontier via Model Merging (Hu et al., Findings 2026)

ACL

• Tiancheng Hu, Benjamin Minixhofer, and Nigel Collier. 2026. Navigating the Alignment-Calibration Trade-off: A Pareto-Superior Frontier via Model Merging. In Findings of the Association for Computational Linguistics: ACL 2026, pages 42405–42422, San Diego, California, United States. Association for Computational Linguistics.