This is a test upload. Not the final dataset.

The final version will include: Cleaned and anonymized Facebook comments Fine-grained annotation labels (SubLabel) Binary labels (BiLabel: Clean vs HOF) Annotation metadata (annotator ID, date) Post-level metadata (page, post, post date, URL, hub, target) This TEST upload is used solely to verify: file structure metadata fields DOI generation versioning workflow on Zenodo Do NOT cite or use this test version for research.A complete and validated dataset will be released in the official version. Short test keywords test dataset placeholder Maghrebi Arabic NLP hate speech (test)