Shi Shuo Xin Yu

该数据集包含了来自《世说新语》的双语数据，涵盖了36个章节，主要用于古文到现代汉语的翻译以及人名识别。数据集按章节划分，提供了原文、译文以及双语对照三种格式，是评估ChatGPT在古文翻译和人名识别性能方面的全面资源。该数据集规模包含3923个句子，其中有300个句子是手动标注用于人名识别的。任务重点在于古文到现代汉语的翻译以及人名识别。This dataset contains bilingual parallel data sourced from *A New Account of the Tales of the World*, covering 36 chapters. It is primarily intended for classical Chinese to modern Chinese translation and personal name recognition. Organized by chapters, it provides three formats: original classical Chinese text, modern Chinese translated text, and bilingual parallel text. With a total of 3,923 sentences, 300 of which are manually annotated for personal name recognition tasks, this dataset serves as a comprehensive resource for evaluating ChatGPT's performance in classical Chinese translation and personal name recognition. The core tasks of this dataset focus on classical Chinese to modern Chinese translation and personal name recognition.

搜集汇总

背景与挑战

背景概述

该数据集是一个大规模的文言文（古文）-现代文平行语料库，涵盖327本古籍原文和97本古籍的句子级双语句对，总计约97万句对。数据经过归一化编辑距离和长度比算法进行对齐，并保留原文顺序，适用于文言文翻译及相关自然语言处理任务。