THOS: A Benchmark Dataset for Targeted Hate and Offensive Speech

Saad Almohaimeed, Saleh Almohaimeed, Ashfaq Ali Shafin, Bogdan Carbunar, Ladislau Bölöni
11/11/2023
cs.CLcs.AI

No official code found. Best implementation: mohaimeed/THOS (0 stars, last commit 2025-06-12).

Abstract

Detecting harmful content on social media, such as Twitter, is made difficult by the fact that the seemingly simple yes/no classification conceals a significant amount of complexity. Unfortunately, while several datasets have been collected for training classifiers in hate and offensive speech, there is a scarcity of datasets labeled with a finer granularity of target classes and specific targets. In this paper, we introduce THOS, a dataset of 8.3k tweets manually labeled with fine-grained annotations about the target of the message. We demonstrate that this dataset makes it feasible to train classifiers, based on Large Language Models, to perform classification at this level of granularity.

View on arXivView PDF

Code Implementations(1)

mohaimeed/THOSCommunity100%

THOS: A Benchmark Dataset for Targeted Hate and Offensive Speech

02Jun 22, 20231 years agoNOASSERTION

Similar Papers

Hate Speech Detection in Limited Data Contexts using Synthetic Data Generation

87% match2023

AlexU-AIC at Arabic Hate Speech 2022: Contrast to Classify

87% match2022

DoDo Learning: DOmain-DemOgraphic Transfer in Language Models for Detecting Abuse Targeted at Public Figures

86% match2023

Developing Linguistic Patterns to Mitigate Inherent Human Bias in Offensive Language Detection

86% match1 repo72023

HateGPT: Unleashing GPT-3.5 Turbo to Combat Hate Speech on X

86% match2024

Offensive Hebrew Corpus and Detection using BERT

86% match1 repo42023

Cite this paper

@article{almohaimeed2023thos,
  title  = {THOS: A Benchmark Dataset for Targeted Hate and Offensive Speech},
  author = {Saad Almohaimeed and Saleh Almohaimeed and Ashfaq Ali Shafin and Bogdan Carbunar and Ladislau Bölöni},
  year   = {2023},
  eprint = {2311.06446},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2311.06446v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2311.06446.svg)](https://codeofpaper.com/paper/2311.06446)

Discussion