Replete-AI/Everything_Instruct_Multilingual

Everything Instruct (Multilingual Edition)是一个大规模的多语言指令格式数据集，旨在提升开源AI的语言模型能力。数据集包含多种主题，如科学、社交媒体、常识、多语言、烹饪、写作、医学、历史、法律、角色扮演、新闻、编码、数学、函数调用和通用指令。数据集支持15种语言，并且每个条目的最大上下文窗口为78,451个标记。数据集未经过滤，除非另有调整，否则模型不会拒绝任何请求。

Everything Instruct (Multilingual Edition) is a massive alpaca instruct formatted dataset consisting of a wide variety of topics meant to bring LLMs to the next level in open source AI. This dataset is fully uncensored and supports multiple languages including English, Russian, Chinese, Korean, Urdu, Latin, Arabic, German, Spanish, French, Hindi, Italian, Japanese, Dutch, and Portuguese. The maximum token context window is 78,451 tokens per line. The dataset includes categories such as science, social media, general knowledge, multilingual content, cooking, writing, medicine, history, law, role-play, news, coding, math, function calling, and general instructions.