Yuanpu Cao

TruthFlow: Truthful LLM Generation via Representation Flow Correction

AdvI2I: Adversarial Image Attack on Image-to-Image Diffusion models

Securing Multimodal Large Language Models: Defending Against Jailbreak Attacks with Adversarial Tuning

AdvI2I: Adversarial Image Attack on Image-to-Image Diffusion Models

Tackling the Data Heterogeneity in Asynchronous Federated Learning with Cached Update Calibration

Personalized Steering of Large Language Models: Versatile Steering Vectors Through Bi-directional Preference Optimization

Defending Against Alignment-Breaking Attacks via Robustly Aligned LLM