The advent of large language models (LLMs) and their capabilities in diverse downstream tasks have attracted considerable adversarial attention, leading to various attack strategies. Among notable ...