Saved in:
Bibliographic Details
Main Authors: de Kock, Christine, Riabi, Arij, Talat, Zeerak, Schlichtkrull, Michael Sejr, Madhyastha, Pranava, Hovy, Ed
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2506.05635
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909640878456832
author de Kock, Christine
Riabi, Arij
Talat, Zeerak
Schlichtkrull, Michael Sejr
Madhyastha, Pranava
Hovy, Ed
author_facet de Kock, Christine
Riabi, Arij
Talat, Zeerak
Schlichtkrull, Michael Sejr
Madhyastha, Pranava
Hovy, Ed
contents Extremist groups develop complex in-group language, also referred to as cryptolects, to exclude or mislead outsiders. We investigate the ability of current language technologies to detect and interpret the cryptolects of two online extremist platforms. Evaluating eight models across six tasks, our results indicate that general purpose LLMs cannot consistently detect or decode extremist language. However, performance can be significantly improved by domain adaptation and specialised prompting techniques. These results provide important insights to inform the development and deployment of automated moderation technologies. We further develop and release novel labelled and unlabelled datasets, including 19.4M posts from extremist platforms and lexicons validated by human experts.
format Preprint
id arxiv_https___arxiv_org_abs_2506_05635
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle IYKYK: Using language models to decode extremist cryptolects
de Kock, Christine
Riabi, Arij
Talat, Zeerak
Schlichtkrull, Michael Sejr
Madhyastha, Pranava
Hovy, Ed
Computation and Language
Extremist groups develop complex in-group language, also referred to as cryptolects, to exclude or mislead outsiders. We investigate the ability of current language technologies to detect and interpret the cryptolects of two online extremist platforms. Evaluating eight models across six tasks, our results indicate that general purpose LLMs cannot consistently detect or decode extremist language. However, performance can be significantly improved by domain adaptation and specialised prompting techniques. These results provide important insights to inform the development and deployment of automated moderation technologies. We further develop and release novel labelled and unlabelled datasets, including 19.4M posts from extremist platforms and lexicons validated by human experts.
title IYKYK: Using language models to decode extremist cryptolects
topic Computation and Language
url https://arxiv.org/abs/2506.05635