Annotated news comments categorized for hate speech detection in Portuguese. The collection captures user-generated text from the Brazilian web to identify offensive language within a specific cultural context.
Use Cases
- Train a hate speech classifier using the comment text and annotation labels.
- Analyze the prevalence of specific offensive terms within the Brazilian news comment corpus.
- Fine-tune large language models for toxicity detection in the Portuguese language.
Strengths
- Contains user-generated news comments sourced from the Brazilian web.
- Includes manual annotations for hate speech detection.
- Focuses exclusively on the Portuguese language.