Yige Li

6

Papers

53

Total Citations

Papers (6)

Memory Injection Attacks on LLM Agents via Query-Only Interaction

NeurIPS 2025arXiv

BlueSuffix: Reinforced Blue Teaming for Vision-Language Models Against Jailbreak Attacks

CROW: Eliminating Backdoors from Large Language Models via Internal Consistency Regularization

Anyattack: Towards Large-scale Self-supervised Adversarial Attacks on Vision-language Models

Backdoor Token Unlearning: Exposing and Defending Backdoors in Pretrained Language Models

Anti-Backdoor Learning: Training Clean Models on Poisoned Data