BaiJia
收藏资源简介:
BaiJia是一个由中国北京邮电大学创建的大规模历史角色扮演数据集,旨在为大语言模型提供低资源的历史角色扮演数据。该数据集包含19281个中国历史人物的信息,涵盖了唐、宋、元、明、清五个朝代。数据集的内容包括人物的传记、文学作品、家庭关系、历史事件等,数据来源广泛,包括历史文献、古籍、艺术作品、民间传说和口述传统。数据集的创建过程包括从多个来源收集人物简历、生成对话以及构建评估问题。该数据集的应用领域主要是增强大语言模型在历史角色扮演任务中的表现,旨在解决历史文本碎片化和多模态数据整合的挑战。
BaiJia is a large-scale historical role-playing dataset developed by Beijing University of Posts and Telecommunications (BUPT) in China, which aims to provide low-resource historical role-playing data for large language models (LLMs). This dataset contains information on 19,281 historical Chinese figures, covering the Tang, Song, Yuan, Ming and Qing dynasties. The dataset content includes biographies, literary works, family relationships, historical events and other relevant details of these figures. The data sources are extensive, covering historical documents, ancient books, artistic works, folklores and oral traditions. The dataset construction process includes collecting the biographical resumes of the figures from multiple sources, generating dialogues, and constructing evaluation questions. Its main application is to enhance the performance of large language models in historical role-playing tasks, aiming to solve the challenges of fragmented historical texts and multimodal data integration.

- 1BaiJia: A Large Scale Role-Playing Agent Corpus of Chinese Historical Charcaters北京邮电大学 · 2024年



