Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMs

Documenti analoghi

• Eight Methods to Evaluate Robust Unlearning in LLMs
  di: Lynch, Aengus, et al.
  Pubblicazione: (2024)
• Defending Against Unforeseen Failure Modes with Latent Adversarial Training
  di: Casper, Stephen, et al.
  Pubblicazione: (2024)
• Randomness, Not Representation: The Unreliability of Evaluating Cultural Alignment in LLMs
  di: Khan, Ariba, et al.
  Pubblicazione: (2025)
• Pitfalls of Evidence-Based AI Policy
  di: Casper, Stephen, et al.
  Pubblicazione: (2025)
• Mechanistic Unlearning: Robust Knowledge Unlearning and Editing via Mechanistic Localization
  di: Guo, Phillip, et al.
  Pubblicazione: (2024)