Minhui Xue
Papers
1
Total Citations
3
H-Index
1
About
Dr. Minhui Xue is a leading researcher at the intersection of artificial intelligence security and privacy, with a particular focus on the vulnerabilities of large language models (LLMs). Her most cited work, "The Philosopher's Stone: Trojaning Plugins of Large Language Models" (2023), has garnered 3 citations and represents a groundbreaking contribution to the field. In this study, Dr. Xue and her colleagues exposed a critical security flaw in open-source LLMs: the ability to inject Trojan attacks through low-rank adapters (LoRAs), which are commonly used to fine-tune models for specialized tasks without expensive hardware. This research fundamentally challenges the assumption that open-source LLMs are inherently safer than their proprietary counterparts, revealing how seemingly benign plugins can be weaponized to manipulate model behavior. Dr. Xue's work has significant implications for the deployment of LLMs in sensitive applications, from healthcare to finance. Her contributions are essential reading for anyone concerned with the security and trustworthiness of modern AI systems, establishing her as a pivotal voice in the ongoing dialogue about AI safety and adversarial robustness.
Research Focus
Key Achievements
Top Papers
- 1The Philosopher's Stone: Trojaning Plugins of Large Language Models3 citations · 2023