Haokun Chen

Reasoning Self-Evaluation via Trajectory Dynamics Modeling

FedNano: Toward Lightweight Federated Tuning for Pretrained Multimodal Large Language Models

Bag of Tricks for Subverting Reasoning-based Safety Guardrails

My Answer Is NOT 'Fair': Mitigating Gender and Race Bias in Vision-Language Models via Fair and Biased Residuals

Visual Instruction Tuning with 500x Fewer Parameters through Modality Linear Representation-Steering

FedBiP: Heterogeneous One-Shot Federated Learning with Personalized Latent Diffusion Models