tlizzo/cTOFU

该数据集包含多个配置，主要针对英语（en）和西班牙语（es）的问答数据，用于机器学习和自然语言处理任务。数据集包括问题、答案、扰动答案（如列表形式的扰动答案）、语言信息（当前语言和源语言）、来源数据集和子集信息、是否机器翻译的标志以及翻译方法。配置变体涉及遗忘（forget）、保留（retain）、保留扰动（retain_perturbed）、真实作者（real_authors）、世界事实（world_facts）、完整（full）和保留部分（holdout）等，可能用于模型训练、评估或数据增强场景，例如模拟数据遗忘、多语言翻译或问答系统测试。每个配置有训练分割，包含不同数量的示例和字节大小，但未提供具体内容描述或应用背景。This dataset includes multiple configurations primarily focused on English (en) and Spanish (es) question-answer data for machine learning and natural language processing tasks. It contains features such as question, answer, perturbed answer (e.g., as a list of perturbed answers), language information (current language and source language), source dataset and subset details, a flag for machine translation, and translation method. Configuration variants involve forget, retain, retain_perturbed, real_authors, world_facts, full, and holdout, likely used for model training, evaluation, or data augmentation scenarios, such as simulating data forgetting, multilingual translation, or question-answering system testing. Each configuration has a train split with varying numbers of examples and byte sizes, but no specific content description or application context is provided.