Abstract Large Language Models (LLMs) are increasingly employed in numerous
applications. It is hence important to ensure that their ethical standard aligns with humans' …

Recent literature has shown that LLMs are vulnerable to backdoor attacks, where malicious
attackers inject a secret token sequence (ie, trigger) into training prompts and enforce their …

Despite prior safety alignment efforts, mainstream LLMs can still generate harmful and
unethical content when subjected to jailbreaking attacks. Existing jailbreaking methods fall …

Large Language Models (LLMs) are susceptible to generating harmful content when
prompted with carefully crafted inputs, a vulnerability known as LLM jailbreaking. As LLMs …

Large Language Models (LLMs) have become integral to many applications, with system
prompts serving as a key mechanism to regulate model behavior and ensure ethical outputs …

Modern large language model (LLM) developers typically conduct a safety alignment to
prevent an LLM from generating unethical or harmful content. Recent studies have …

Recent studies developed jailbreaking attacks, which construct jailbreaking prompts
to``fool''LLMs into responding to harmful questions. Early-stage jailbreaking attacks require …

Large Language Models (LLMs) have become increasingly impactful across various
domains, including coding and data analysis. However, their widespread adoption has …

Jailbreaking attacks can effectively manipulate open-source large language models (LLMs)
to produce harmful responses. Nevertheless, these attacks exhibit limited transferability …