Our paper titled Instruction Backdoor Attacks Against Customized LLMs was accepted by Usenix Security 2024.