Safety Alignment Depth in Large Language Models: A Markov Chain Perspective

Documenti analoghi

• The Great Contradiction Showdown: How Jailbreak and Stealth Wrestle in Vision-Language Models?
  di: Kao, Ching-Chia, et al.
  Pubblicazione: (2024)
• Safe LoRA: the Silver Lining of Reducing Safety Risks when Fine-tuning Large Language Models
  di: Hsu, Chia-Yi, et al.
  Pubblicazione: (2024)
• Defending Against Repetitive Backdoor Attacks on Semi-supervised Learning through Lens of Rate-Distortion-Perception Trade-off
  di: Lee, Cheng-Yi, et al.
  Pubblicazione: (2024)
• Harmless Yet Harmful: Neutral Prompting Attacks for Stealthy Hallucination Steering in Agent Skills
  di: Hsu, Chia-Yi, et al.
  Pubblicazione: (2026)
• Counter-Current Learning: A Biologically Plausible Dual Network Approach for Deep Learning
  di: Kao, Chia-Hsiang, et al.
  Pubblicazione: (2024)