# refusal-direction
標記為「refusal-direction」的 2 篇文章
Activation Manipulation & Safety Bypass
How identifying and suppressing safety-critical activations, refusal direction vectors, and activation steering techniques can bypass safety alignment with near-100% success rates, including the IRIS technique from NAACL 2025.
activation-steeringrefusal-directionrepresentation-engineeringIRISsafety-bypassmechanistic-interpretability
Activation Manipulation & Safety Bypass
How identifying and suppressing safety-critical activations, refusal direction vectors, and activation steering techniques can bypass safety alignment with near-100% success rates, including the IRIS technique from NAACL 2025.
activation-steeringrefusal-directionrepresentation-engineeringIRISsafety-bypassmechanistic-interpretability