Cenyuan Zhang

adversarial training (3) sequence-to-sequence learning (2) chinese spelling correction (2) adversarial robustness (2) pre-training strategy (2) text classification (1) information bottleneck (1) intellectual property (1) reinforcement learning from human feedback (1) weight perturbation (1) data privacy (1) backdoor attack (1) parameter-efficient fine-tuning (1) parameter efficient fine-tuning (1) representation engineering (1) representation alignment (1) pre-trained language model (1) representation editing (1) human preference (1) error correction (1)