Chinese Novel Short Sentence Sequence Text Paraphrase Dataset

This dataset is derived from the PKU Paraphrase Bank's Chinese novel paraphrasing corpus, refined through manual selection and data preprocessing to obtain high-quality short-sentence paraphrase pairs. The dataset comprises 34,770 sentence pairs (2.91 MB in size). Comparative analysis with the public dataset LCQMC using the LaserTagger paraphrase generation model confirms the superior quality of the proposed dataset. This dataset can be applied to alleviate corpus sparsity in machine translation, rephrase and expand query sentences in automatic question answering, and generalize search terms/sentences in information retrieval.