<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="thonet-etal-2026-ready">
    <titleInfo>
        <title>Ready to Speak: Aligning LLMs for TTS-Friendly Text Generation</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Thibaut</namePart>
        <namePart type="family">Thonet</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jos</namePart>
        <namePart type="family">Rozen</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Laurent</namePart>
        <namePart type="family">Besacier</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>Current Large Language Models (LLMs) are primarily optimized for written text, often producing outputs that are grammatically correct and helpful yet poorly suited for spoken delivery via Text-to-Speech (TTS). In this work, we study how to make LLMs natively generate TTS-friendly text, which we frame as a preference alignment problem: instead of relying on downstream rewriting modules, we directly align LLMs to generate text optimized for spoken delivery. We introduce two preference datasets spanning different target domains, CORA and Recipe, which contain paired TTS-friendly and TTS-unfriendly responses. We further propose an evaluation suite combining a pattern-based heuristic metric, a TTS→ASR evaluation pipeline, and a MUSHRA listening study with human judges. Our experiments compare the recently proposed Feature-aware Sampling and Tuning (FaST) framework – leveraging interpretable features instead of a black-box reward model – against an array of alignment baselines on the TTS-friendly generation task. Notably, we found that FaST achieves the best overall tradeoff between TTS-friendliness and helpfulness across various settings. We also identified a strong correlation between our different metrics, highlighting the ability to reliably assess TTS-friendliness via an efficient heuristic.</abstract>
    <identifier type="citekey">thonet-etal-2026-ready</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.248/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>5650</start>
            <end>5676</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Ready to Speak: Aligning LLMs for TTS-Friendly Text Generation
%A Thonet, Thibaut
%A Rozen, Jos
%A Besacier, Laurent
%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 thonet-etal-2026-ready
%X Current Large Language Models (LLMs) are primarily optimized for written text, often producing outputs that are grammatically correct and helpful yet poorly suited for spoken delivery via Text-to-Speech (TTS). In this work, we study how to make LLMs natively generate TTS-friendly text, which we frame as a preference alignment problem: instead of relying on downstream rewriting modules, we directly align LLMs to generate text optimized for spoken delivery. We introduce two preference datasets spanning different target domains, CORA and Recipe, which contain paired TTS-friendly and TTS-unfriendly responses. We further propose an evaluation suite combining a pattern-based heuristic metric, a TTS→ASR evaluation pipeline, and a MUSHRA listening study with human judges. Our experiments compare the recently proposed Feature-aware Sampling and Tuning (FaST) framework – leveraging interpretable features instead of a black-box reward model – against an array of alignment baselines on the TTS-friendly generation task. Notably, we found that FaST achieves the best overall tradeoff between TTS-friendliness and helpfulness across various settings. We also identified a strong correlation between our different metrics, highlighting the ability to reliably assess TTS-friendliness via an efficient heuristic.
%U https://aclanthology.org/2026.emnlp-main.248/
%P 5650-5676

Markdown (Informal)

[Ready to Speak: Aligning LLMs for TTS-Friendly Text Generation](https://aclanthology.org/2026.emnlp-main.248/) (Thonet et al., EMNLP 2026)

• Ready to Speak: Aligning LLMs for TTS-Friendly Text Generation (Thonet et al., EMNLP 2026)

ACL

• Thibaut Thonet, Jos Rozen, and Laurent Besacier. 2026. Ready to Speak: Aligning LLMs for TTS-Friendly Text Generation. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 5650–5676, Budapest, Hungary. Association for Computational Linguistics.