The Geometry of Refusal in Large Language Models: Concept Cones and Representational Independence

Documenti analoghi

• REINFORCE Adversarial Attacks on Large Language Models: An Adaptive, Distributional, and Semantic Objective
  di: Geisler, Simon, et al.
  Pubblicazione: (2025)
• Attacking Large Language Models with Projected Gradient Descent
  di: Geisler, Simon, et al.
  Pubblicazione: (2024)
• Consistency Training while Mitigating Obfuscation via Rate Matching
  di: Imran, Sohaib, et al.
  Pubblicazione: (2026)
• RepIt: Steering Language Models with Concept-Specific Refusal Vectors
  di: Siu, Vincent, et al.
  Pubblicazione: (2025)
• Solving an Open Problem in Theoretical Physics using AI-Assisted Discovery
  di: Brenner, Michael P., et al.
  Pubblicazione: (2026)