HateSpeechCorpus
HateSpeechCorpus is a dataset collected from Twitter, consisting of 3003 tweets. It has been meticulously annotated by three speech-language pathology graduate students, ensuring high-quality labeling of hate speech. This dataset is invaluable for researchers and practitioners working on hate speech detection and natural language processing.
Source: Twitter
Length: 3003 tweets
Annotators: Three speech-language pathology graduate students
Paper
For a detailed investigation of annotator bias in LLMs for hate speech detection, refer to the associated paper: Investigating Annotator Bias in Large Language Models for Hate Speech Detection.
Citation
If you use this dataset in your research, please cite the following paper:
@misc{das2024investigating, title={Investigating Annotator Bias in Large Language Models for Hate Speech Detection}, author={Amit Das and Zheng Zhang and Fatemeh Jamshidi and Vinija Jain and Aman Chadha and Nilanjana Raychawdhary and Mary Sandage and Lauramarie Pope and Gerry Dozier and Cheryl Seals}, year={2024}, eprint={2406.11109}, archivePrefix={arXiv}, primaryClass={cs.CL} }
Load the dataset
df = pd.read_csv('HateSpeechCorpus.csv')
Display the first few rows
print(df.head())