Maple LeafBURST
  • Read Later
  • Top Stories
  • latest
  • bc
  • business
  • canada
  • economy
  • education
  • entertainment
  • health
  • lifestyle
  • politics
  • research
  • science
  • sports
  • technology
  • world
  • technology

A small number of samples can poison LLMs of any size

Hacker News
Thu 12m read
Comments
www.anthropic.com

Related stories

  • technology
  • #AI
  • #BizIT
  • #AIResearch
  • #AISecurity
  • #AIVulnerabilities
  • #AlanTuringInstitute
  • #Anthropic
  • #BackdoorAttacks
  • #DataPoisoning
  • #FineTuning
  • #LLMSecurity
  • #MachineLearning
  • #ModelSafety
  • #Pretraining
  • #TrainingData
  • #UKAISecurityInstitute

AI models can acquire backdoors from surprisingly few malicious documents

Ars Technica
Thu 3m read
No bubbles, just news