This dataset contains four classes of text samples designed for research on machine-generated text detection, stylometric transformations, and LLM-based text post-processing.
Each text sample belongs to one of the following categories:
hw — Human-written: sourced directly from the publicly available human-authored dataset on Zenodo.
mw — Machine-written: generated by prompting LLMs with the first sentence of a human-written text.
hw_mp — Human-written… See the full description on the dataset page:
https://huggingface.co/datasets/tanishy7777/IITGN_GPT_dataset.
