Papers by Andrzej Prałat
BAN-PL: A Polish Dataset of Banned Harmful and Offensive Content from Wykop.pl Web Service (2024.lrec-main)
Copied to clipboard
Anna Kolos, Inez Okulska, Kinga Głąbińska, Agnieszka Karlinska, Emilia Wisnios, Paweł Ellerik, Andrzej Prałat
| Challenge: | a new dataset of offensive social media content for the Polish language is presented to address this gap . access to accurate and non-synthetic datasets of social media is limited for low-resource languages . |
| Approach: | They present a new open dataset of offensive social media content for the Polish language . authors propose to make the dataset publicly available to improve access . |
| Outcome: | The proposed dataset includes 691,662 posts and comments from the Polish Reddit . the authors describe the dataset and apply it to real-life content moderation processes . |