1 paper · 1 filter
Pia Pachinger, Janis Goldzycher, Anna Maria Planitzer +3
Model interpretability in toxicity detection greatly profits from token-level annotations. However, currently such annotations are only available in English. We introduce a dataset…