Lu Lin

Understanding and Rectifying Safety Perception Distortion in VLMs

JoPA: Explaining Large Language Model's Generation via Joint Prompt Attribution

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

Mitigating Graph Covariate Shift via Score-based Out-of-distribution Augmentation

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

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

Backdoor Contrastive Learning via Bi-level Trigger Optimization

Defending Against Alignment-Breaking Attacks via Robustly Aligned LLM

On the Safety of Open-Sourced Large Language Models: Does Alignment Really Prevent Them From Being Misused?