NataliaVanetik/OffensiveLanguageResearchLab

Contains offensive language datasets for under-resourced languages # Emotional and Offensive Language Research Lab Hub Welcome to the central hub for datasets and research from the Offensive Language Research Lab. Our work is dedicated to advancing the understanding and detection of various forms of problematic language online, including offensive speech, hate speech, propaganda, and negative campaigning. This document provides a comprehensive and organized collection of the datasets we have developed, along with links to the corresponding research papers and repositories. We aim to make these resources accessible to the broader research community to foster collaboration and further innovation in this critical area of Natural Language Processing. ### Quick Overview of Datasets | *Dataset* | *Language* | *Primary Task* | *Size* | |--|--|--|--| | **OLaH-2000** | Hebrew | Offensive Language Detection | 2,000 comments | | **OLaH-5000** | Hebrew | Offensive Language Detection | 5,000 comments | | **OLaH-450** | Hebrew | Fine-grained Offense Taxonomy Classification | 450 instances | | **OLaH-500** | Hebrew | Text Detoxification | 450 instances | | **OLaA** | Arabic | Offensive Language Detection | 9,000 comments | | **FARAD** | Arabic | Offensive Language Classification | 6,459 instances | | **FARAD-500** | Arabic | Fine-grained Offense Taxonomy Classification | 500 instances | | **TONIC** | Hebrew | Detecting negative campaigns in the context of Israeli municipal elections | 2,632 comments | | **FTR** | French | Detecting racist language in French | 2,856 tweets | | **Propaganda-Ru** | Russian | Propaganda Detection | 11,320 documents | | **heDetox** | Hebrew | Hebrew detoxification dataset | 600 texts | | **reFARAD-500** | Arabic | Fine-Grained Offensive Language Classification | 500 instances | | **HeMotion** | Hebrew | Plutchik-based emotion classification | 670 texts | | **ArEkmanPlus** | Arabic | Ekman-based emotion classification | ~31K tex …