上海人工智能实验室联合商汤科技、香港中文大学和复旦大学，近日正式发布了新一代大语言模型书生·浦语2.0，并将其核心算法开源。这一版本的研发主旨是回归语言建模的本质，通过提升训练数据的质量与信息密度，从而显著提高模型的语言理解及生成能力。
书生·浦语2.0支持高达200K token的上下文长度，能够处理一次性约30万汉字的输入内容，使其在处理长篇文本信息时，具备了在海量数据中精准提取关键信息的能力，类比于在茫茫大海中捞针。这一特性对于需要深度理解和分析大规模文本数据的科研、媒体和商业应用场景，具有极高的实用价值。
据悉，书生·浦语2.0的开源将促进全球AI研究社区的交流与合作，进一步推动人工智能技术的发展。它的开放不仅有助于提升国内外的语言模型研究水平，而且对促进中文自然语言处理技术的进步具有里程碑意义。
英文翻译：
Shanghai AI Lab Unveils Open Source Scholar·Puyu 2.0
Keywords: AI Lab, Scholar·Puyu 2.0, Language Modeling
News content:
The Shanghai Artificial Intelligence Laboratory, in collaboration with SenseTime Technology, the Chinese University of Hong Kong, and Fudan University, has recently officially launched the new-generation large language model Scholar·Puyu 2.0 and open-sourced its core algorithm. The development philosophy of this version is to return to the essence of language modeling, which aims to enhance the model’s language understanding and generation capabilities by improving the quality and information density of the training data.
Scholar·Puyu 2.0 supports a context length of up to 200K tokens and can handle input content of about 300,000 Chinese characters at once, enabling it to accurately extract key information from large-scale text data, akin to “fishing for a needle in a haystack”. This feature holds significant practical value for scientific research, media, and business applications that require in-depth understanding and analysis of vast amounts of textual data.
It is understood that the open-source nature of Scholar·Puyu 2.0 will foster communication and collaboration within the global AI research community, further propelling the development of artificial intelligence technology. Its openness is not only beneficial for enhancing the research level of language models at home and abroad but also marks a milestone in promoting the advancement of Chinese natural language processing technology.
【来源】https://www.thepaper.cn/newsDetail_forward_26040295
Views: 0