Yifan Yang, Qiao Jin, Furong Huang, Zhiyong Lu. "Adversarial prompt and fine-tuning attacks threaten medical large language models." Nature Communications (2025). https://doi.org/10.1038/s41467-025-64062-1