Tutoring Dialog Dataset
收藏资源简介:
Tutoring Dialog Dataset是由九州大学信息科学与电气工程研究生院开发的,旨在模拟学生与导师之间的真实对话,特别是在阅读理解任务中。该数据集包含63个阅读理解问题,涵盖了从2年级到5年级的不同难度和主题的文本。数据集的创建过程包括收集阅读理解工作表、生成合成对话以及通过人类评估者进行评估和增强。该数据集主要用于优化大型语言模型(LLMs)在教育领域的应用,特别是作为一对一辅导工具,旨在提高学生的阅读理解和批判性思维能力。
The Tutoring Dialog Dataset was developed by the Graduate School of Information Science and Electrical Engineering, Kyushu University, to simulate realistic dialogues between students and tutors, with a specific focus on reading comprehension tasks. The dataset includes 63 reading comprehension questions, paired with texts spanning various difficulty levels and academic topics for grades 2 through 5. The development of this dataset involved collecting reading comprehension worksheets, generating synthetic dialogues, and subsequently evaluating and refining the data through human evaluators. This dataset is primarily intended to optimize the deployment of Large Language Models (LLMs) in educational scenarios, particularly as a one-on-one tutoring tool, with the objective of enhancing students' reading comprehension and critical thinking skills.




