Trelis/yourbench-ait

该数据集包含了多种配置，每个配置针对不同的文本处理任务。具体包括：
- chunked配置：包含文档的ID、文本、文件名、文件大小、摘要、分块信息及分块评价指标。
- ingested配置：包含文档的ID、文本、文件名和文件大小。
- lighteval配置：包含问题、答案、问题类别、难度评估、引用、文档ID、分块ID、问题和文档文本。
- multi_hop_questions配置：包含文档ID、源分块ID、问题、自我答案、难度评估、问题类型、生成模型、思考过程和引用。
- single_shot_questions配置：包含分块ID、文档ID、问题、自我答案、难度评估、问题类型、生成模型、思考过程和引用。
- summarized配置：包含文档的ID、文本、文件名、文件大小和摘要。

The dataset consists of multiple configurations, each tailored for different text processing tasks. It includes:
- chunked configuration: contains document ID, text, filename, file size, summary, chunk information, and chunk metric evaluations.
- ingested configuration: contains document ID, text, filename, and file size.
- lighteval configuration: contains questions, answers, question categories, difficulty assessments, citations, document ID, chunk IDs, and document text.
- multi_hop_questions configuration: contains document ID, source chunk IDs, questions, self-answers, estimated difficulty, self-assessed question type, generating model, thought process, and citations.
- single_shot_questions configuration: contains chunk ID, document ID, question, self-answer, estimated difficulty, self-assessed question type, generating model, thought process, and citations.
- summarized configuration: contains document ID, text, filename, file size, and summary.