This dataset contains train/validation/test splits for binary AI-generated text detection.
It is built from sources whose dataset-level licenses were checked as permissive or
public-domain-compatible.
text: essay text.
label: 0 for human-written text, 1 for AI-generated text.
source_dataset: upstream dataset identifier.
source_detail: source label retained from the upstream data.
source_license: row-level… See the full description on the dataset page:
https://huggingface.co/datasets/sinatras/isogram-ai-text-detection-splits.
