bigstupidhats/dynasample_train

该数据集包含多语言对话数据，涵盖了阿拉伯语、德语、英语、西班牙语、印地语、俄语、斯瓦希里语和中文等多种语言。数据集中的每个对话都包含了指令、输出、语言、语言代码、注释类型、用户ID等信息。此外，数据集还包含了对话的毒性检测和内容审核信息，如openai_moderation和detoxify_moderation字段，用于评估对话内容的安全性。数据集按语言划分，每个语言版本的数据量和示例数量也有所不同。This dataset contains multilingual dialogue data, covering languages such as Arabic, German, English, Spanish, Hindi, Russian, Swahili, and Chinese. Each dialogue in the dataset includes information such as instruction, output, language, language code, annotation type, and user ID. Additionally, the dataset includes toxicity detection and content moderation information, such as the openai_moderation and detoxify_moderation fields, to assess the safety of the dialogue content. The dataset is divided by language, with varying data volumes and example counts for each language version.