Popular repositories Loading
-
-
SplitLLM-main
SplitLLM-main Public该项目揭示了基于分割框架的大型语言模型私有微调漏洞,提出双向半白盒重构攻击(BiSR),结合预训练权重先验与双向优化,可有效重构私有数据,即使在扰动防御下仍表现强劲。
-
SplitLLM
SplitLLM Public大语言模型分割学习仿真框架,支持多种数据重构攻击与防护设计 采用攻防一致的设置,支持7种攻击方法(DLG/TAG/LAMP/BiSR/SIP/EIA/SMA)和3种防御机制(高斯噪声/DxPrivacy/Inf2Guard)
Something went wrong, please refresh the page to try again.
If the problem persists, check the GitHub status page or contact support.
If the problem persists, check the GitHub status page or contact support.