LLM-Generated Feedback Dataset
收藏资源简介:
本数据集由卡内基梅隆大学的研究团队创建,旨在研究大型语言模型(LLM)生成的解释性反馈对学习的影响。数据集包含来自885名辅导学习者的2,648个课程完成记录,涵盖了七个基于场景的辅导培训课程。研究通过比较不同组别学习者在后测中的表现,探讨了LLM反馈对学习的效果。数据集提供了对LLM反馈有效性的实证支持,为开放性任务的学习改进提供了低成本且可扩展的方法。
This dataset was developed by a research team at Carnegie Mellon University to investigate the impact of explanatory feedback generated by Large Language Models (LLMs) on learning. It contains 2,648 course completion records from 885 tutored learners, spanning seven scenario-based tutoring training courses. The study examined the effects of LLM feedback on learning by comparing the post-test performance of different learner groups. This dataset offers empirical evidence supporting the effectiveness of LLM-generated feedback, providing a low-cost and scalable method for improving learning in open-ended tasks.




