Self and Cross-Model Distillation for LLMs: Effective Methods for Refusal Pattern Alignment

Documenti analoghi

• A Cross-Language Investigation into Jailbreak Attacks in Large Language Models
  di: Li, Jie, et al.
  Pubblicazione: (2024)
• Do Reasoning LLMs Refuse What They Infer in Long Contexts?
  di: Fu, Yu, et al.
  Pubblicazione: (2026)
• Beyond Surface Alignment: Rebuilding LLMs Safety Mechanism via Probabilistically Ablating Refusal Direction
  di: Xie, Yuanbo, et al.
  Pubblicazione: (2025)
• The Art of (Mis)alignment: How Fine-Tuning Methods Effectively Misalign and Realign LLMs in Post-Training
  di: Zhang, Rui, et al.
  Pubblicazione: (2026)
• From Threat to Tool: Leveraging Refusal-Aware Injection Attacks for Safety Alignment
  di: Chae, Kyubyung, et al.
  Pubblicazione: (2025)