Privacy Risks of General-Purpose Language Modelsseminar @ Tencent Meeting, Tsinghua University, Beijing Presenters: Mingxuan Liu .Date: 14:00-15:10pm November 12, 2020Abstract:通用语言模型中的文本嵌入从纯文本中捕获了大量的敏感信息,可以通过逆向工程来披露受害者的敏感信息,以便进一步骚扰。Paper SlidesShare on Twitter Facebook LinkedIn Previous Next