上海人工智能实验室与商汤科技联合香港中文大学和复旦大学正式发布新一代大语言模型书生·浦语2.0。浦语2.0支持200K token的上下文,一次性接收并处理约30万汉字的输入,准确提取关键信息。其核心在于通过提高语料质量及信息密度,实现模型语言建模能力获得质的提升。浦语2.0可实现长文本“大海捞针”,致力于回归语言建模的本质。
Title: PuYu 2.0 open sourced, AI language ability improved
Keywords: PuYu 2.0, language model, information extraction
News content: Shanghai AI Lab and SenseTime, together with CUHK and Fudan University, officially released PuYu 2.0, the next generation large language model. PuYu 2.0 supports 200K token context and processes 300,000 Chinese characters input in one go, accurately extracting key information. Its core is to qualitatively improve model’s language modeling capability by enhancing corpus quality and information density. PuYu 2.0 can find a needle in an ocean of text, committed to returning to the essence of language modeling.
【来源】https://www.thepaper.cn/newsDetail_forward_26040295
Views: 1