Latent Adversarial Training Improves the Representation of Refusal

Documenti analoghi

• Refusal in LLMs is an Affine Function
  di: Marshall, Thomas, et al.
  Pubblicazione: (2024)
• Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMs
  di: Sheshadri, Abhay, et al.
  Pubblicazione: (2024)
• Dynamic Adversarial Fine-Tuning Reorganizes Refusal Geometry
  di: Lan, Wenhao, et al.
  Pubblicazione: (2026)
• Tracing the Dynamics of Refusal: Exploiting Latent Refusal Trajectories for Robust Jailbreak Detection
  di: Hu, Xulin, et al.
  Pubblicazione: (2026)
• Refusal Tokens: A Simple Way to Calibrate Refusals in Large Language Models
  di: Jain, Neel, et al.
  Pubblicazione: (2024)