Deactivating Refusal Triggers: Understanding and Mitigating Overrefusal in Safety Alignment

Documenti analoghi

• No Free Lunch for Defending Against Prefilling Attack by In-Context Learning
  di: Xue, Zhiyu, et al.
  Pubblicazione: (2024)
• Enhancing the Safety of Medical Vision-Language Models by Synthetic Demonstrations
  di: Xue, Zhiyu, et al.
  Pubblicazione: (2025)
• Conflict-Aware Adversarial Training
  di: Xue, Zhiyu, et al.
  Pubblicazione: (2024)
• Understanding and Mitigating Overrefusal in LLMs from an Unveiling Perspective of Safety Decision Boundary
  di: Pan, Licheng, et al.
  Pubblicazione: (2025)
• Inverse Reinforcement Learning by Estimating Expertise of Demonstrators
  di: Beliaev, Mark, et al.
  Pubblicazione: (2024)