大语言模型智能体中长期记忆安全综述:迈向记忆主权 A Survey on the Security of Long-Term Memory in LLM Agents Toward Mnemonic Sovereignty.pdfVIP

  • 2
  • 0
  • 约25.87万字
  • 约 63页
  • 2026-06-03 发布于广东
  • 举报

大语言模型智能体中长期记忆安全综述:迈向记忆主权 A Survey on the Security of Long-Term Memory in LLM Agents Toward Mnemonic Sovereignty.pdf

ASurveyontheSecurityofLong-TermMemoryinLLM

Agents:TowardMnemonicSovereignty

ZEHAOLIN,MemTensor,China

CHUNYULI,MemTensor,China

KAICHEN,MemTensor,China

Researchonthesecurityoflargelanguagemodel(LLM)systemsisshiftingfrom“willthemodelleaktraining

6data”toamoreconsequentialquestion:“cananagentwithpersistent,long-termmemorybecontinuously

2shaped,cross-sessionpoisoned,accessedwithoutauthorization,andpropagatedacrosssharedorganizational

0state?”Recentsurveyshave

您可能关注的文档

文档评论(0)

1亿VIP精品文档

相关文档