Dataset Card for GAHD

Dataset Description

GAHD is a German Adversarial Hate speech Dataset containing 10,996 examples. We collected the dataset via four rounds of Dynamic Adversarial Data Collection and explored various methods of supporting annotators in finding adversarial examples.

• Paper: https://arxiv.org/abs/2403.19559
• Repository: https://github.com/jagol/gahd

Dataset Structure

gahd.csv contains the following columns:

• gahd_id: unique identifier of the entry
• text: text of the entry
• label: 0 = "not-hate speech", 1 = "hate speech"
• round: round in which the entry was created
• split: "train", "dev", or "test"
• contrastive_gahd_id: gahd_id of its contrastive example

gahd_disaggregated.csv contains the following additional columns:

• source:
  • if annotators entered the entry via the Dynabench interface: dynabench
  • if the entry was translated from the Vidgen et al. 2021 dataset: translation
  • if the entry stems from the Leipzit news corpus: news
• model_prediction: label predicted by the target model, 0 or 1
• annotator_id: unique identifier of the annotator that created the entry
• annotator_labels: a string containing a forward slash-separated list of all labels by annotators
• expert_labels: 0 or 1 if an expert annotator annotated the entry, otherwise empty

Citation

When using GAHD, please cite our preprint on Arxiv:

@misc{goldzycher2024improving,
      title={Improving Adversarial Data Collection by Supporting Annotators: Lessons from GAHD, a German Hate Speech Dataset}, 
      author={Janis Goldzycher and Paul Röttger and Gerold Schneider},
      year={2024},
      eprint={2403.19559},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}