metadata
license: openrail++
Dataset formation:
- Filtering Ukrainian tweets so that only tweets containing toxic language remain. Source of Ukrainian data: https://github.com/saganoren/ukr-twi-corpus
- Non-toxic sentences were obtained from a previous dataset of tweets as well as sentences from news and fiction from UD Ukrainian IU: https://universaldependencies.org/treebanks/uk_iu/index.html
- After that, the dataset was split into a train-test-val and all data were balanced both by the toxic/non-toxic criterion and by data source.