Zidi Xiong

backdoor attack (2) conformal prediction (1) probabilistic modeling (1) anomaly detection (1) adversarial machine learning (1) toxicity detection (1) multilingual reasoning (1) clustering approach (1) certified defense (1) large reasoning model (1) unsupervised model detection (1) adversarial target (1) large language model (1) reasoning accuracy (1) language mismatch (1) trustworthiness evaluation (1) x2x attack (1) adversarial robustness (1) thinking trace (1)