cl-nagoya/wikisplit-pp
WikiSplit++ This dataset is the HuggingFace version of WikiSplit++.WikiSplit++ enhances the original WikiSplit by applying two techniques: filtering through NLI classification and sentence-order reversing, which help to remove noise and reduce hallucinations compared to the original WikiSplit.The preprocessed WikiSplit dataset that formed the basis for this can be found here. Usage import datasets as ds dataset: ds.DatasetDict =… See the full description on the dataset page: https://huggingface.co/datasets/cl-nagoya/wikisplit-pp.
387
