Efficient LLM Jailbreak via Adaptive Dense-to-sparse Constrained Optimization This is a draft implementation of our paper [NeurIPS 2024] Efficient LLM Jailbreak via Adaptive Dense-to-sparse Constrained Optimization An official implementation will be ready soon.