@inproceedings{weidinger-etal-2024-star,
title = "{STAR}: {S}ocio{T}echnical Approach to Red Teaming Language Models",
author = "Weidinger, Laura and
Mellor, John F J and
Pegueroles, Bernat Guill{\'e}n and
Marchal, Nahema and
Kumar, Ravin and
Lum, Kristian and
Akbulut, Canfer and
Diaz, Mark and
Bergman, A. Stevie and
Rodriguez, Mikel D. and
Rieser, Verena and
Isaac, William",
editor = "Al-Onaizan, Yaser and
Bansal, Mohit and
Chen, Yun-Nung",
booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2024",
address = "Miami, Florida, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.emnlp-main.1200",
doi = "10.18653/v1/2024.emnlp-main.1200",
pages = "21516--21532",
abstract = "This research introduces STAR, a sociotechnical framework that improves on current best practices for red teaming safety of large language models. STAR makes two key contributions: it enhances steerability by generating parameterised instructions for human red teamers, leading to improved coverage of the risk surface. Parameterised instructions also provide more detailed insights into model failures at no increased cost. Second, STAR improves signal quality by matching demographics to assess harms for specific groups, resulting in more sensitive annotations. STAR further employs a novel step of arbitration to leverage diverse viewpoints and improve label reliability, treating disagreement not as noise but as a valuable contribution to signal quality.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="weidinger-etal-2024-star">
<titleInfo>
<title>STAR: SocioTechnical Approach to Red Teaming Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Laura</namePart>
<namePart type="family">Weidinger</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">John</namePart>
<namePart type="given">F</namePart>
<namePart type="given">J</namePart>
<namePart type="family">Mellor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bernat</namePart>
<namePart type="given">Guillén</namePart>
<namePart type="family">Pegueroles</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nahema</namePart>
<namePart type="family">Marchal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ravin</namePart>
<namePart type="family">Kumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kristian</namePart>
<namePart type="family">Lum</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Canfer</namePart>
<namePart type="family">Akbulut</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mark</namePart>
<namePart type="family">Diaz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">A</namePart>
<namePart type="given">Stevie</namePart>
<namePart type="family">Bergman</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mikel</namePart>
<namePart type="given">D</namePart>
<namePart type="family">Rodriguez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Verena</namePart>
<namePart type="family">Rieser</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">William</namePart>
<namePart type="family">Isaac</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yaser</namePart>
<namePart type="family">Al-Onaizan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohit</namePart>
<namePart type="family">Bansal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yun-Nung</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Miami, Florida, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>This research introduces STAR, a sociotechnical framework that improves on current best practices for red teaming safety of large language models. STAR makes two key contributions: it enhances steerability by generating parameterised instructions for human red teamers, leading to improved coverage of the risk surface. Parameterised instructions also provide more detailed insights into model failures at no increased cost. Second, STAR improves signal quality by matching demographics to assess harms for specific groups, resulting in more sensitive annotations. STAR further employs a novel step of arbitration to leverage diverse viewpoints and improve label reliability, treating disagreement not as noise but as a valuable contribution to signal quality.</abstract>
<identifier type="citekey">weidinger-etal-2024-star</identifier>
<identifier type="doi">10.18653/v1/2024.emnlp-main.1200</identifier>
<location>
<url>https://aclanthology.org/2024.emnlp-main.1200</url>
</location>
<part>
<date>2024-11</date>
<extent unit="page">
<start>21516</start>
<end>21532</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T STAR: SocioTechnical Approach to Red Teaming Language Models
%A Weidinger, Laura
%A Mellor, John F. J.
%A Pegueroles, Bernat Guillén
%A Marchal, Nahema
%A Kumar, Ravin
%A Lum, Kristian
%A Akbulut, Canfer
%A Diaz, Mark
%A Bergman, A. Stevie
%A Rodriguez, Mikel D.
%A Rieser, Verena
%A Isaac, William
%Y Al-Onaizan, Yaser
%Y Bansal, Mohit
%Y Chen, Yun-Nung
%S Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
%D 2024
%8 November
%I Association for Computational Linguistics
%C Miami, Florida, USA
%F weidinger-etal-2024-star
%X This research introduces STAR, a sociotechnical framework that improves on current best practices for red teaming safety of large language models. STAR makes two key contributions: it enhances steerability by generating parameterised instructions for human red teamers, leading to improved coverage of the risk surface. Parameterised instructions also provide more detailed insights into model failures at no increased cost. Second, STAR improves signal quality by matching demographics to assess harms for specific groups, resulting in more sensitive annotations. STAR further employs a novel step of arbitration to leverage diverse viewpoints and improve label reliability, treating disagreement not as noise but as a valuable contribution to signal quality.
%R 10.18653/v1/2024.emnlp-main.1200
%U https://aclanthology.org/2024.emnlp-main.1200
%U https://doi.org/10.18653/v1/2024.emnlp-main.1200
%P 21516-21532
Markdown (Informal)
[STAR: SocioTechnical Approach to Red Teaming Language Models](https://aclanthology.org/2024.emnlp-main.1200) (Weidinger et al., EMNLP 2024)
ACL
- Laura Weidinger, John F J Mellor, Bernat Guillén Pegueroles, Nahema Marchal, Ravin Kumar, Kristian Lum, Canfer Akbulut, Mark Diaz, A. Stevie Bergman, Mikel D. Rodriguez, Verena Rieser, and William Isaac. 2024. STAR: SocioTechnical Approach to Red Teaming Language Models. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 21516–21532, Miami, Florida, USA. Association for Computational Linguistics.