Pinned Loading

1. ToLD-Br ToLD-Br Public

   [AACL 2020] Toxic Language Detection in Social Media for Brazilian Portuguese: New Dataset and Multilingual Analysis

   Jupyter Notebook 51 12

2. PASTEL PASTEL Public

   [EPJ Data Science] PASTEL (Prompted weAk Supervision wiTh crEdibility signaLs) is a weakly supervised approach that leverages large language models to extract credibility signals from web content, …

   Jupyter Notebook 4

3. euvsdisinfo euvsdisinfo Public

   [CIKM 2024] Allows collecting the EUvsDisinfo dataset and reproducing the research experiments presented in the paper 'EUvsDisinfo: A Dataset for Multilingual Detection of Pro-Kremlin Disinformatio…

   Python 2 2

4. persuasion-injection-attack persuasion-injection-attack Public

   [EMNLP 2026 Main] Persuasion techniques as adversarial attacks on automated fact-checking systems.

   Python