Computer Science > Computation and Language
[Submitted on 8 Sep 2023 (v1), last revised 31 Mar 2024 (this version, v4)]
Title:Down the Toxicity Rabbit Hole: A Novel Framework to Bias Audit Large Language Models
View PDF HTML (experimental)Abstract:This paper makes three contributions. First, it presents a generalizable, novel framework dubbed \textit{toxicity rabbit hole} that iteratively elicits toxic content from a wide suite of large language models. Spanning a set of 1,266 identity groups, we first conduct a bias audit of \texttt{PaLM 2} guardrails presenting key insights. Next, we report generalizability across several other models. Through the elicited toxic content, we present a broad analysis with a key emphasis on racism, antisemitism, misogyny, Islamophobia, homophobia, and transphobia. Finally, driven by concrete examples, we discuss potential ramifications.
Submission history
From: Ashiqur Rahman KhudaBukhsh [view email][v1] Fri, 8 Sep 2023 03:59:02 UTC (1,699 KB)
[v2] Mon, 18 Sep 2023 16:56:40 UTC (1,699 KB)
[v3] Sat, 23 Dec 2023 06:54:20 UTC (1,701 KB)
[v4] Sun, 31 Mar 2024 02:24:39 UTC (3,391 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.