@inproceedings{naber-etal-2026-probing,
title = "Probing Discrete Speech Tokens of Spoken Language Models",
author = "Naber, Sven and
Koch, Julia and
Singh, Pranav and
Saponaro, Alberto and
Karagianni, Ioanna and
Vu, Ngoc Thang",
editor = "Piperidis, Stelios and
Bel, N{\'u}ria and
van den Heuvel, Henk and
Ide, Nancy and
Krek, Simon and
Toral, Antonio",
booktitle = "Proceedings of the Fifteenth Language Resources and Evaluation Conference",
month = may,
year = "2026",
address = "Palma de Mallorca, Spain",
publisher = "ELRA Language Resource Association",
url = "https://aclanthology.org/2026.lrec-1.184/",
doi = "10.63317/4d4kttgdydde",
pages = "2344--2354",
abstract = "This paper presents a framework for systematic probing of discrete speech token representations in spoken language models (SLMs). We propose three complementary components: a distributional divergence analysis testing whether an attribute is reflected in token usage, token-based classifiers to quantify recoverability and an attribute-conditioned representation analysis revealing phonetic attribute realizations. As a demonstration we apply these probes to tokenizer outputs and model generations from CosyVoice2 and SparkTTS on LibriTTS-R and VCTK. We find that gender is encoded in their respective tokens but in different forms - the signal is more stable across stages and datasets in CosyVoice2, whereas SparkTTS shows weaker cross-stage consistency and stronger pause/prosody-related effects. Exploratory probes of valence, arousal, and dominance are weaker and less consistent. These results show that discrete speech tokens retain speaker-related information in different ways across architectures and that the proposed framework provides an interpretable basis for comparing token representations across spoken language modeling pipelines."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="naber-etal-2026-probing">
<titleInfo>
<title>Probing Discrete Speech Tokens of Spoken Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sven</namePart>
<namePart type="family">Naber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Julia</namePart>
<namePart type="family">Koch</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pranav</namePart>
<namePart type="family">Singh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alberto</namePart>
<namePart type="family">Saponaro</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ioanna</namePart>
<namePart type="family">Karagianni</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ngoc</namePart>
<namePart type="given">Thang</namePart>
<namePart type="family">Vu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Fifteenth Language Resources and Evaluation Conference</title>
</titleInfo>
<name type="personal">
<namePart type="given">Stelios</namePart>
<namePart type="family">Piperidis</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Núria</namePart>
<namePart type="family">Bel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Henk</namePart>
<namePart type="family">van den Heuvel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nancy</namePart>
<namePart type="family">Ide</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simon</namePart>
<namePart type="family">Krek</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Antonio</namePart>
<namePart type="family">Toral</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resource Association</publisher>
<place>
<placeTerm type="text">Palma de Mallorca, Spain</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>This paper presents a framework for systematic probing of discrete speech token representations in spoken language models (SLMs). We propose three complementary components: a distributional divergence analysis testing whether an attribute is reflected in token usage, token-based classifiers to quantify recoverability and an attribute-conditioned representation analysis revealing phonetic attribute realizations. As a demonstration we apply these probes to tokenizer outputs and model generations from CosyVoice2 and SparkTTS on LibriTTS-R and VCTK. We find that gender is encoded in their respective tokens but in different forms - the signal is more stable across stages and datasets in CosyVoice2, whereas SparkTTS shows weaker cross-stage consistency and stronger pause/prosody-related effects. Exploratory probes of valence, arousal, and dominance are weaker and less consistent. These results show that discrete speech tokens retain speaker-related information in different ways across architectures and that the proposed framework provides an interpretable basis for comparing token representations across spoken language modeling pipelines.</abstract>
<identifier type="citekey">naber-etal-2026-probing</identifier>
<identifier type="doi">10.63317/4d4kttgdydde</identifier>
<location>
<url>https://aclanthology.org/2026.lrec-1.184/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>2344</start>
<end>2354</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Probing Discrete Speech Tokens of Spoken Language Models
%A Naber, Sven
%A Koch, Julia
%A Singh, Pranav
%A Saponaro, Alberto
%A Karagianni, Ioanna
%A Vu, Ngoc Thang
%Y Piperidis, Stelios
%Y Bel, Núria
%Y van den Heuvel, Henk
%Y Ide, Nancy
%Y Krek, Simon
%Y Toral, Antonio
%S Proceedings of the Fifteenth Language Resources and Evaluation Conference
%D 2026
%8 May
%I ELRA Language Resource Association
%C Palma de Mallorca, Spain
%F naber-etal-2026-probing
%X This paper presents a framework for systematic probing of discrete speech token representations in spoken language models (SLMs). We propose three complementary components: a distributional divergence analysis testing whether an attribute is reflected in token usage, token-based classifiers to quantify recoverability and an attribute-conditioned representation analysis revealing phonetic attribute realizations. As a demonstration we apply these probes to tokenizer outputs and model generations from CosyVoice2 and SparkTTS on LibriTTS-R and VCTK. We find that gender is encoded in their respective tokens but in different forms - the signal is more stable across stages and datasets in CosyVoice2, whereas SparkTTS shows weaker cross-stage consistency and stronger pause/prosody-related effects. Exploratory probes of valence, arousal, and dominance are weaker and less consistent. These results show that discrete speech tokens retain speaker-related information in different ways across architectures and that the proposed framework provides an interpretable basis for comparing token representations across spoken language modeling pipelines.
%R 10.63317/4d4kttgdydde
%U https://aclanthology.org/2026.lrec-1.184/
%U https://doi.org/10.63317/4d4kttgdydde
%P 2344-2354
Markdown (Informal)
[Probing Discrete Speech Tokens of Spoken Language Models](https://aclanthology.org/2026.lrec-1.184/) (Naber et al., LREC 2026)
ACL
- Sven Naber, Julia Koch, Pranav Singh, Alberto Saponaro, Ioanna Karagianni, and Ngoc Thang Vu. 2026. Probing Discrete Speech Tokens of Spoken Language Models. In Proceedings of the Fifteenth Language Resources and Evaluation Conference, pages 2344–2354, Palma de Mallorca, Spain. ELRA Language Resource Association.