yourbench-testing/test-custom-schema-simple

该数据集名为Test Custom Schema Simple，是通过YourBench（v0.6.0）开源框架生成的，用于从文档集合中生成特定领域的基准测试。数据集包含三个配置：chunked、ingested和single_shot_questions，分别对应不同的数据处理步骤和特征。ingestion步骤读取原始文档并转换为标准化markdown；chunking步骤将文本分割为基于token的单跳和多跳块；single_shot_question_generation步骤使用LLM为每个块生成独立的问答对。数据集旨在为特定领域的基准测试提供支持。

This dataset, named Test Custom Schema Simple, was generated using YourBench (v0.6.0), an open-source framework for generating domain-specific benchmarks from document collections. The dataset includes three configurations: chunked, ingested, and single_shot_questions, each corresponding to different data processing steps and features. The ingestion step reads raw source documents and converts them to normalized markdown; the chunking step splits texts into token-based single-hop and multi-hop chunks; the single_shot_question_generation step generates standalone question-answer pairs per chunk using LLM. The dataset is designed to support domain-specific benchmarking.