sumuks/celerai_education_v2

该数据集包含了多种配置，用于不同的NLP任务。配置包括：分段（chunked）、已摄入（ingested）、轻量评估（lighteval）、多跳问题（multi_hop_questions）、单次问答（single_shot_questions）和摘要（summarized）。每个配置都包含文档的ID、文本、文件名等特征，并根据配置的不同，提供文本摘要、分段信息、问题生成模型等信息。数据集适用于文本摘要、问答系统等任务。

The dataset consists of multiple configurations for different NLP tasks, including chunked, ingested, lighteval, multi_hop_questions, single_shot_questions, and summarized. Each configuration contains features such as document ID, text, filename, and depending on the configuration, it provides text summaries, chunk information, question generation models, and more. The dataset is suitable for tasks like text summarization and question-answering systems.