<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="sternberg-etal-2026-interleaved">
    <titleInfo>
        <title>Interleaved Speech Language Models Latently Work In Text</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Talia</namePart>
        <namePart type="family">Sternberg</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Gallil</namePart>
        <namePart type="family">Maimon</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yossi</namePart>
        <namePart type="family">Adi</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>Speech language models (SLMs) increasingly combine speech and text, often by interleaving their tokens within a single sequence. Yet how these two modalities interact in the model’s latent space remains unclear. In this work, we analyze interleaved speech–text LMs from different model families and training configurations using three complementary methods. We reveal that these models pass through an implicit latent transcription phase in which the text token matching the spoken word becomes decodable in intermediate layers, despite not being trained for speech recognition. This phenomenon occurs in diverse, natural speech, and intermediate representations also encode likely text continuations. We further show that implicit transcription emerges most when combining text-LM pretraining and speech–text interleaving, and that its prevalence is positively associated with spoken factual-knowledge retrieval. Our analysis sheds light on the internal interaction between speech and text modalities in interleaved SLMs.</abstract>
    <identifier type="citekey">sternberg-etal-2026-interleaved</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.2398/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>51527</start>
            <end>51555</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Interleaved Speech Language Models Latently Work In Text
%A Sternberg, Talia
%A Maimon, Gallil
%A Adi, Yossi
%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 sternberg-etal-2026-interleaved
%X Speech language models (SLMs) increasingly combine speech and text, often by interleaving their tokens within a single sequence. Yet how these two modalities interact in the model’s latent space remains unclear. In this work, we analyze interleaved speech–text LMs from different model families and training configurations using three complementary methods. We reveal that these models pass through an implicit latent transcription phase in which the text token matching the spoken word becomes decodable in intermediate layers, despite not being trained for speech recognition. This phenomenon occurs in diverse, natural speech, and intermediate representations also encode likely text continuations. We further show that implicit transcription emerges most when combining text-LM pretraining and speech–text interleaving, and that its prevalence is positively associated with spoken factual-knowledge retrieval. Our analysis sheds light on the internal interaction between speech and text modalities in interleaved SLMs.
%U https://aclanthology.org/2026.findings-emnlp.2398/
%P 51527-51555

Markdown (Informal)

[Interleaved Speech Language Models Latently Work In Text](https://aclanthology.org/2026.findings-emnlp.2398/) (Sternberg et al., Findings 2026)

• Interleaved Speech Language Models Latently Work In Text (Sternberg et al., Findings 2026)

ACL

• Talia Sternberg, Gallil Maimon, and Yossi Adi. 2026. Interleaved Speech Language Models Latently Work In Text. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 51527–51555, Budapest, Hungary. Association for Computational Linguistics.