Best-of-Venom: Attacking RLHF by Injecting Poisoned Preference Data

Documenti analoghi

• Enhancing Prompt Injection Attacks to LLMs via Poisoning Alignment
  di: Shao, Zedian, et al.
  Pubblicazione: (2024)
• Universal Jailbreak Backdoors from Poisoned Human Feedback
  di: Rando, Javier, et al.
  Pubblicazione: (2023)
• Special Characters Attack: Toward Scalable Training Data Extraction From Large Language Models
  di: Bai, Yang, et al.
  Pubblicazione: (2024)
• Machine Unlearning Fails to Remove Data Poisoning Attacks
  di: Pawelczyk, Martin, et al.
  Pubblicazione: (2024)
• BadReward: Clean-Label Poisoning of Reward Models in Text-to-Image RLHF
  di: Duan, Kaiwen, et al.
  Pubblicazione: (2025)