lintonsui/low-resource-lang-research

该数据集包含多语言的指令和对应的响应，每条数据包括唯一标识符、语言、指令文本、响应文本、来源、原始语言、数据集类型、主题、难度、风格、角色和长度等字段。数据分为实验组（treatment）和对照组（control），各有训练集和验证集，可能用于评估指令遵循模型或对话系统的实验性研究。This dataset contains multilingual instructions and corresponding responses. Each entry includes fields such as id, language, instruction, response, source, original_language, dataset_type, topic, difficulty, style, persona, and length. The data is split into treatment and control groups, each with training and validation sets, likely for experimental studies on instruction-following models or dialogue systems.