Unveiling the Hate: Generating Faithful and Plausible Explanations for Implicit and Subtle Hate Speech Detection - Archive ouverte HAL
Communication Dans Un Congrès Année : 2024

Unveiling the Hate: Generating Faithful and Plausible Explanations for Implicit and Subtle Hate Speech Detection

Résumé

In today’s digital age, the huge amount of abusive content and hate speech on social media platforms presents a significant chal- lenge. Natural Language Processing (NLP) methods have focused on detecting explicit forms of hate speech, often overlooking more nuanced and implicit instances. To address this gap, our paper aims to enhance the detection and understanding of implicit and subtle hate speech. More precisely, we propose a comprehensive approach combining prompt con- struction, free-text generation, few-shot learning, and fine-tuning to gen- erate explanations for hate speech classification, with the goal of provid- ing more context for content moderators to unveil the actual nature of a message on social media.
Fichier principal
Vignette du fichier
NLDB_24-8.pdf (421.29 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Commentaire This preprint corresponds with the author’s version of the paper.

Dates et versions

hal-04658110 , version 1 (22-07-2024)
hal-04658110 , version 2 (30-07-2024)

Licence

Identifiants

  • HAL Id : hal-04658110 , version 2

Citer

Greta Damo, Nicolás Benjamín Ocampo, Elena Cabrio, Serena Villata. Unveiling the Hate: Generating Faithful and Plausible Explanations for Implicit and Subtle Hate Speech Detection. NLDB 2024 - 29th International Conference on Natural Language & Information Systems, Jun 2024, Torino, Italy. ⟨hal-04658110v2⟩
148 Consultations
83 Téléchargements

Partager

More