Exploring the Adversarial Capabilities of Large Language Models

arXiv:2402.09132v1 Announce Type: cross
Abstract: The proliferation of large language models (LLMs) has sparked widespread and general interest due to their strong language generation capabilities, offering great potential for both industry and research. While previous research delved into the security and privacy issues of LLMs, the extent to which these models can exhibit adversarial behavior remains largely unexplored. Addressing this gap, we investigate whether common publicly available LLMs have inherent capabilities to perturb text samples to fool safety measures, so-called …

abstract adversarial arxiv capabilities cs.ai cs.lg general industry language language generation language models large language large language models llms privacy research security security and privacy type