LARGO: Latent Adversarial Reflection through Gradient Optimization for Jailbreaking LLMs

Open in new window