Replication Package for the Paper "Digital Guardians: Can GPT-4, Perspective API, and Moderation API reliably detect hate speech in reader comments of German online newspapers?"

This replication package contains the text samples and model predictions used for the evaluation in the paper "Digital Guardians: Can GPT-4, Perspective API, and Moderation API reliably detect hate speech in reader comments of German online newspapers?" presented at FRUCT 2025.samples_and_model_predictions.csv contains: Label – The human annotation: 1 (hate speech) or 0 (no hate speech)Text – The calssified text sample 1., 2., and 3. run prediction of each evaluated model reannotated_samples.csv contains:Original Label – The initial label as provided in samples_and_model_predictions.csvFinal Label – The corrected label obtained through reannotation, determined by majority vote among three independent human annotators.Text – The calssified text sample