<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="liu-etal-2026-reliability">
    <titleInfo>
        <title>Reliability-Calibrated Guidance for Few-Step Categorical Flow-Map Text Generation</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Zihao</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Hongchen</namePart>
        <namePart type="family">Wu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Xiaochang</namePart>
        <namePart type="family">Fang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Hongxuan</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Huaxiang</namePart>
        <namePart type="family">Zhang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jiande</namePart>
        <namePart type="family">Sun</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
    </relatedItem>
    <abstract>Few-step flow-map guidance provides an efficient way to steer generation without retraining, but its updates are defined in a continuous or relaxed state space. In categorical text generation, however, the final output is determined by simplex-constrained token distributions and hard decoding. Under low NFE, where each update induces a relatively large finite displacement, this continuous-to-discrete mismatch can make terminal reward guidance unreliable, causing overshoot, reward–quality conflict, saturation, or poor transfer from relaxed reward gains to decoded tokens.\par We present Reliability-Calibrated Flow Guidance (RCFG), a sampling-time reliability-calibration framework for terminal reward pullback in few-step categorical flow maps. Rather than replacing the underlying guidance primitive, RCFG calibrates its magnitude and direction according to transition scale, reward saturation, local reward–quality conflict, and categorical endpoint structure. Experiments on LM1B and Text8 show more favorable aggregate control–likelihood operating points after calibration: on LM1B, RCFG maintains or improves verifier Reward and Target while reducing NLL and PPL relative to terminal reward-pullback guidance, while Text8 shows a modest aggregate improvement in lexical-control balance. Code and reproduction instructions are available at: https://github.com/ZionLau/RCFG</abstract>
    <identifier type="citekey">liu-etal-2026-reliability</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.1995/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>42517</start>
            <end>42546</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Reliability-Calibrated Guidance for Few-Step Categorical Flow-Map Text Generation
%A Liu, Zihao
%A Wu, Hongchen
%A Fang, Xiaochang
%A Li, Hongxuan
%A Zhang, Huaxiang
%A Sun, Jiande
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F liu-etal-2026-reliability
%X Few-step flow-map guidance provides an efficient way to steer generation without retraining, but its updates are defined in a continuous or relaxed state space. In categorical text generation, however, the final output is determined by simplex-constrained token distributions and hard decoding. Under low NFE, where each update induces a relatively large finite displacement, this continuous-to-discrete mismatch can make terminal reward guidance unreliable, causing overshoot, reward–quality conflict, saturation, or poor transfer from relaxed reward gains to decoded tokens.\par We present Reliability-Calibrated Flow Guidance (RCFG), a sampling-time reliability-calibration framework for terminal reward pullback in few-step categorical flow maps. Rather than replacing the underlying guidance primitive, RCFG calibrates its magnitude and direction according to transition scale, reward saturation, local reward–quality conflict, and categorical endpoint structure. Experiments on LM1B and Text8 show more favorable aggregate control–likelihood operating points after calibration: on LM1B, RCFG maintains or improves verifier Reward and Target while reducing NLL and PPL relative to terminal reward-pullback guidance, while Text8 shows a modest aggregate improvement in lexical-control balance. Code and reproduction instructions are available at: https://github.com/ZionLau/RCFG
%U https://aclanthology.org/2026.findings-emnlp.1995/
%P 42517-42546

Markdown (Informal)

[Reliability-Calibrated Guidance for Few-Step Categorical Flow-Map Text Generation](https://aclanthology.org/2026.findings-emnlp.1995/) (Liu et al., Findings 2026)

• Reliability-Calibrated Guidance for Few-Step Categorical Flow-Map Text Generation (Liu et al., Findings 2026)

ACL

• Zihao Liu, Hongchen Wu, Xiaochang Fang, Hongxuan Li, Huaxiang Zhang, and Jiande Sun. 2026. Reliability-Calibrated Guidance for Few-Step Categorical Flow-Map Text Generation. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 42517–42546, Budapest, Hungary. Association for Computational Linguistics.