FigSIM: A Dataset for Fine-grained Suicide Severity and Figurative Language in Suicide Memes

Liuliu Chen, Elise R. Carrotte, Brian E. Chapman, Jo Robinson, Mike Conway · 2026

Read FigSIM: A Dataset for Fine-grained Suicide Severity and Figurative Language in Suicide Memes by Liuliu Chen, Elise R. Carrotte, Brian E. Chapman, Jo Robinson, Mike Conway (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-02523.txt.

Description

Chen and co-authors release FigSIM, a dataset for the study of suicide memes. Such memes voice suicide-related thoughts or comment on the topic. They grow more common on social media. Their traits stay poorly understood, and they can cause harm. Moderation tools need annotated data to learn from. That data has been missing until now.

FigSIM holds 1,049 annotated memes. Each meme is tagged on three axes. Tags cover fine-grained suicide severity levels. They cover figurative devices like metaphors. They cover suicide-related content such as method depiction. The team benchmarks 16 unimodal and multimodal models. Tasks span figurative language, severity, and content detection. Models underpredict high severity levels. The bias is worst for figurative memes. The dataset and its splits are public. The authors frame suicide memes as a distinct challenge for models and moderators alike.

Catalog record

CREATORLiuliu Chen, Elise R. Carrotte, Brian E. Chapman, Jo Robinson, Mike Conway
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE51 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.02523

Browse the full library · About the Tower of Babel