Lifan Yuan

adversarial defense (2) textual adversarial attack (2) natural language processing (2) adversarial attack (2) large language model (2) model security (1) few-shot learning (1) benchmark evaluation (1) gradient-based optimization (1) text classification (1) uncertainty quantification (1) model calibration (1) data augmentation (1) in-context learning (1) adversarial robustness (1) instruction tuning (1) language model alignment (1) preference optimization (1) distribution shift (1) masked language model (1)