yourbench-testing/childrens-books-test
收藏Hugging Face2025-12-14 更新2025-12-20 收录
下载链接:
https://hf-mirror.com/datasets/yourbench-testing/childrens-books-test
下载链接
链接失效反馈官方服务:
资源简介:
该数据集名为儿童书籍测试,是一个基于儿童书籍的测试数据集,通过YourBench框架生成。数据集包含多个配置,如分块(chunked)、原始数据(ingested)、轻量评估准备(prepared_lighteval)、单次问题生成(single_shot_questions)和摘要(summarized)。每个配置具有不同的特征和分割。数据集的生成过程包括多个步骤:数据读取与标准化、分层摘要、文本分块以及独立问答对生成。README还提供了用于复现数据集的详细配置信息。
This dataset, named Childrens Books Test, is a test dataset based on childrens books, generated using the YourBench framework. It includes multiple configurations such as chunked, ingested, prepared_lighteval, single_shot_questions, and summarized, each with distinct features and splits. The dataset generation process involves several steps: ingestion (reading and normalizing source documents), summarization (hierarchical summarization), chunking (splitting texts into single-hop and multi-hop chunks), and single_shot_question_generation (generating standalone question-answer pairs). The README also provides configuration details for reproducing the dataset.
提供机构:
yourbench-testing



