This repository hosts a special version of the well-known OffensEval-TR 2020 dataset.
This version splits the official training dataset into a slightly smaller training dataset and a development set. Purpose of this special version is to have a more consistent dataset for benchmarks.
Homepage: offensive-turkish
Paper: A Corpus of Turkish Offensive Language on Social Media
Point of Contact: Çağrı Çöltekin
Original Dataset on… See the full description on the dataset page:
https://huggingface.co/datasets/Jokesta75/offenseval2020_tr.