Offensive Language Identification Dataset for Brazilian Portuguese (OLID-BR) is a dataset with multi-task annotations for the detection of offensive language.
The current version (v1.0) contains 7,943 (extendable to 13,538) comments from different sources, including social media (YouTube and Twitter) and related datasets.
OLID-BR contains a collection of annotated sentences in Brazilian Portuguese using an annotation model that encompasses the following levels:
Offensive… See the full description on the dataset page:
https://huggingface.co/datasets/dougtrajano/olid-br.
