Refusal Behavior in Large Language Models: A Nonlinear Perspective

Fabian Hildebrandt, Andreas Maier, Patrick Krauss +1

Refusal behavior in large language models (LLMs) enables them to decline responding to harmful, unethical, or inappropriate prompts, ensuring alignment with ethical standards. This…