<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="singh-yannakoudakis-2026-gradient">
    <titleInfo>
        <title>Gradient-Free Test-Time Adaptation for Few-Shot Online Clustering with Large Language Models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Avyav</namePart>
        <namePart type="given">Kumar</namePart>
        <namePart type="family">Singh</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Helen</namePart>
        <namePart type="family">Yannakoudakis</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>We study few-shot online clustering (FSOC), where an unlabelled stream must be assigned to a dynamically growing cluster inventory using only a few labelled support examples for currently observed classes. We propose a two-component test-time gradient-free framework. First, we adapt Wasserstein Generative Flow Networks (wGFN) as an amortised LLM-based policy for in-distribution assignment and out-of-distribution (OOD) rejection. Unlike prior work that trains separate policies for each evaluation class set, we show that a single policy trained on a few classes generalises to novel support classes at test time without parameter updates and further replace fixed rejection thresholds with a Dirichlet evidence model over candidate class assignments. Second, rejected OOD examples enter an oracle-mediated cluster-formation stage. Under shuffled stream orderings, this stage can create noisy provisional support sets that corrupt later assignments. To address this, we introduce a von Mises–Fisher Dirichlet Process Mixture Model warm start to provide oracle-independent cluster structure over unit-normalised text embeddings before oracle-mediated formation. This improves clustering quality, in some settings by \sim 2\times, even with small LLMs (&lt;1B parameters). To our knowledge, this is the first FSOC framework to couple amortised LLM-based assignment with Bayesian non-parametric warm-starting, enabling growing cluster inventories without data augmentation, novel-class parameter updates or predefined cluster counts.</abstract>
    <identifier type="citekey">singh-yannakoudakis-2026-gradient</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.373/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>7865</start>
            <end>7891</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Gradient-Free Test-Time Adaptation for Few-Shot Online Clustering with Large Language Models
%A Singh, Avyav Kumar
%A Yannakoudakis, Helen
%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 singh-yannakoudakis-2026-gradient
%X We study few-shot online clustering (FSOC), where an unlabelled stream must be assigned to a dynamically growing cluster inventory using only a few labelled support examples for currently observed classes. We propose a two-component test-time gradient-free framework. First, we adapt Wasserstein Generative Flow Networks (wGFN) as an amortised LLM-based policy for in-distribution assignment and out-of-distribution (OOD) rejection. Unlike prior work that trains separate policies for each evaluation class set, we show that a single policy trained on a few classes generalises to novel support classes at test time without parameter updates and further replace fixed rejection thresholds with a Dirichlet evidence model over candidate class assignments. Second, rejected OOD examples enter an oracle-mediated cluster-formation stage. Under shuffled stream orderings, this stage can create noisy provisional support sets that corrupt later assignments. To address this, we introduce a von Mises–Fisher Dirichlet Process Mixture Model warm start to provide oracle-independent cluster structure over unit-normalised text embeddings before oracle-mediated formation. This improves clustering quality, in some settings by \sim 2\times, even with small LLMs (<1B parameters). To our knowledge, this is the first FSOC framework to couple amortised LLM-based assignment with Bayesian non-parametric warm-starting, enabling growing cluster inventories without data augmentation, novel-class parameter updates or predefined cluster counts.
%U https://aclanthology.org/2026.findings-emnlp.373/
%P 7865-7891

Markdown (Informal)

[Gradient-Free Test-Time Adaptation for Few-Shot Online Clustering with Large Language Models](https://aclanthology.org/2026.findings-emnlp.373/) (Singh & Yannakoudakis, Findings 2026)

• Gradient-Free Test-Time Adaptation for Few-Shot Online Clustering with Large Language Models (Singh & Yannakoudakis, Findings 2026)

ACL

• Avyav Kumar Singh and Helen Yannakoudakis. 2026. Gradient-Free Test-Time Adaptation for Few-Shot Online Clustering with Large Language Models. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 7865–7891, Budapest, Hungary. Association for Computational Linguistics.