Publikation:

generAItor : Tree-in-the-Loop Text Generation for Language Model Explainability and Adaptation

Lade...
Vorschaubild

Dateien

Spinner_2-17c7rv7yw33dp0.PDF
Spinner_2-17c7rv7yw33dp0.PDFGröße: 2.85 MBDownloads: 64

Datum

2024

Herausgeber:innen

Kontakt

ISSN der Zeitschrift

Electronic ISSN

ISBN

Bibliografische Daten

Verlag

Schriftenreihe

Auflagebezeichnung

DOI (zitierfähiger Link)
ArXiv-ID

Internationale Patentnummer

Link zur Lizenz
oops

Angaben zur Forschungsförderung

Deutsche Forschungsgemeinschaft (DFG): 390829875
Deutsche Forschungsgemeinschaft (DFG): 240796339 (FOR 2111)

Projekt

Open Access-Veröffentlichung
Open Access Green
Core Facility der Universität Konstanz

Gesperrt bis

Titel in einer weiteren Sprache

Publikationstyp
Zeitschriftenartikel
Publikationsstatus
Published

Erschienen in

ACM Transactions on Interactive Intelligent Systems. Association for Computing Machinery (ACM). 2024, 14(2), 14. ISSN 2160-6455. eISSN 2160-6463. Verfügbar unter: doi: 10.1145/3652028

Zusammenfassung

Large language models (LLMs) are widely deployed in various downstream tasks, e.g., auto-completion, aided writing, or chat-based text generation. However, the considered output candidates of the underlying search algorithm are under-explored and under-explained. We tackle this shortcoming by proposing a tree-in-the-loop approach, where a visual representation of the beam search tree is the central component for analyzing, explaining, and adapting the generated outputs. To support these tasks, we present generAItor, a visual analytics technique, augmenting the central beam search tree with various task-specific widgets, providing targeted visualizations and interaction possibilities. Our approach allows interactions on multiple levels and offers an iterative pipeline that encompasses generating, exploring, and comparing output candidates, as well as fine-tuning the model based on adapted data. Our case study shows that our tool generates new insights in gender bias analysis beyond state-of-the-art template-based methods. Additionally, we demonstrate the applicability of our approach in a qualitative user study. Finally, we quantitatively evaluate the adaptability of the model to few samples, as occurring in text-generation use cases.

Zusammenfassung in einer weiteren Sprache

Fachgebiet (DDC)
004 Informatik

Schlagwörter

Konferenz

Rezension
undefined / . - undefined, undefined

Forschungsvorhaben

Organisationseinheiten

Zeitschriftenheft

Zugehörige Datensätze in KOPS

Zitieren

ISO 690SPINNER, Thilo, Rebecca KEHLBECK, Rita SEVASTJANOVA, Tobias STÄHLE, Daniel A. KEIM, Oliver DEUSSEN, Mennatallah EL-ASSADY, 2024. generAItor : Tree-in-the-Loop Text Generation for Language Model Explainability and Adaptation. In: ACM Transactions on Interactive Intelligent Systems. Association for Computing Machinery (ACM). 2024, 14(2), 14. ISSN 2160-6455. eISSN 2160-6463. Verfügbar unter: doi: 10.1145/3652028
BibTex
@article{Spinner2024-06-30gener-69728,
  year={2024},
  doi={10.1145/3652028},
  title={generAItor : Tree-in-the-Loop Text Generation for Language Model Explainability and Adaptation},
  number={2},
  volume={14},
  issn={2160-6455},
  journal={ACM Transactions on Interactive Intelligent Systems},
  author={Spinner, Thilo and Kehlbeck, Rebecca and Sevastjanova, Rita and Stähle, Tobias and Keim, Daniel A. and Deussen, Oliver and El-Assady, Mennatallah},
  note={Article Number: 14}
}
RDF
<rdf:RDF
    xmlns:dcterms="http://purl.org/dc/terms/"
    xmlns:dc="http://purl.org/dc/elements/1.1/"
    xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
    xmlns:bibo="http://purl.org/ontology/bibo/"
    xmlns:dspace="http://digital-repositories.org/ontologies/dspace/0.1.0#"
    xmlns:foaf="http://xmlns.com/foaf/0.1/"
    xmlns:void="http://rdfs.org/ns/void#"
    xmlns:xsd="http://www.w3.org/2001/XMLSchema#" > 
  <rdf:Description rdf:about="https://kops.uni-konstanz.de/server/rdf/resource/123456789/69728">
    <dcterms:isPartOf rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dc:creator>Stähle, Tobias</dc:creator>
    <dc:date rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2024-04-02T10:49:17Z</dc:date>
    <dc:creator>Kehlbeck, Rebecca</dc:creator>
    <dc:creator>El-Assady, Mennatallah</dc:creator>
    <dspace:isPartOfCollection rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dc:language>eng</dc:language>
    <dspace:isPartOfCollection rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/43615"/>
    <dc:creator>Keim, Daniel A.</dc:creator>
    <dc:contributor>El-Assady, Mennatallah</dc:contributor>
    <foaf:homepage rdf:resource="http://localhost:8080/"/>
    <dc:contributor>Deussen, Oliver</dc:contributor>
    <dcterms:issued>2024-06-30</dcterms:issued>
    <dc:contributor>Stähle, Tobias</dc:contributor>
    <dc:contributor>Keim, Daniel A.</dc:contributor>
    <dcterms:isPartOf rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/43615"/>
    <dcterms:available rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2024-04-02T10:49:17Z</dcterms:available>
    <dc:contributor>Kehlbeck, Rebecca</dc:contributor>
    <dcterms:abstract>Large language models (LLMs) are widely deployed in various downstream tasks, e.g., auto-completion, aided writing, or chat-based text generation. However, the considered output candidates of the underlying search algorithm are under-explored and under-explained. We tackle this shortcoming by proposing a tree-in-the-loop approach, where a visual representation of the beam search tree is the central component for analyzing, explaining, and adapting the generated outputs. To support these tasks, we present generAItor, a visual analytics technique, augmenting the central beam search tree with various task-specific widgets, providing targeted visualizations and interaction possibilities. Our approach allows interactions on multiple levels and offers an iterative pipeline that encompasses generating, exploring, and comparing output candidates, as well as fine-tuning the model based on adapted data. Our case study shows that our tool generates new insights in gender bias analysis beyond state-of-the-art template-based methods. Additionally, we demonstrate the applicability of our approach in a qualitative user study. Finally, we quantitatively evaluate the adaptability of the model to few samples, as occurring in text-generation use cases.</dcterms:abstract>
    <dc:contributor>Sevastjanova, Rita</dc:contributor>
    <dc:creator>Deussen, Oliver</dc:creator>
    <dspace:hasBitstream rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/69728/1/Spinner_2-17c7rv7yw33dp0.PDF"/>
    <dc:contributor>Spinner, Thilo</dc:contributor>
    <dc:creator>Sevastjanova, Rita</dc:creator>
    <void:sparqlEndpoint rdf:resource="http://localhost/fuseki/dspace/sparql"/>
    <dcterms:title>generAItor : Tree-in-the-Loop Text Generation for Language Model Explainability and Adaptation</dcterms:title>
    <dc:creator>Spinner, Thilo</dc:creator>
    <dcterms:hasPart rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/69728/1/Spinner_2-17c7rv7yw33dp0.PDF"/>
    <bibo:uri rdf:resource="https://kops.uni-konstanz.de/handle/123456789/69728"/>
  </rdf:Description>
</rdf:RDF>

Interner Vermerk

xmlui.Submission.submit.DescribeStep.inputForms.label.kops_note_fromSubmitter

Kontakt
URL der Originalveröffentl.

Prüfdatum der URL

Prüfungsdatum der Dissertation

Finanzierungsart

Kommentar zur Publikation

Allianzlizenz
Corresponding Authors der Uni Konstanz vorhanden
Internationale Co-Autor:innen
Universitätsbibliographie
Ja
Begutachtet
Ja
Diese Publikation teilen