Zhaorun Chen

ARMs: Adaptive Red-Teaming Agent against Multimodal Models with Plug-and-Play Attacks

Scaling Agent Learning via Experience Synthesis

Guarding a Needle in the Haystack: A Real-Time Policy-Following Streaming Video Guardrail

The Devil behind the mask: An emergent safety vulnerability of Diffusion LLMs

Grounding and Enhancing Informativeness and Utility in Dataset Distillation

Cer-Eval: Certifiable and Cost-Efficient Evaluation Framework for LLMs

RedCodeAgent: Automatic Red-teaming Agent against Diverse Code Agents

Beyond Reward Hacking: Causal Rewards for Large Language Model Alignment

Agent Learning via Early Experience

From EduVisBench to EduVisAgent: A Benchmark and Multi-Agent Framework for Reasoning-Driven Pedagogical Visualization

Towards Principled Dataset Distillation: A Spectral Distribution Perspective

Multi-Sample Preference Optimization for Generative Model Alignment

GTA: Graph Theory Agent and Benchmark for Algorithmic Graph Reasoning with LLMs

SocialMaze: A Benchmark for Evaluating Social Reasoning in Large Language Models

Federated Agent Reinforcement Learning

AgentXploit: End-to-End Red-Teaming for AI Agents Powdered by Multi-Agent Systems

SAFEVISION: Efficient Image Guardrail with Robust Policy Adherence and Explainability

MMIE: Massive Multimodal Interleaved Comprehension Benchmark for Large Vision-Language Models

MJ-Video: Benchmarking and Rewarding Video Generation with Fine-Grained Video Preference

SafeWatch: An Efficient Safety-Policy Following Video Guardrail Model with Transparent Explanations

MMDT: Decoding the Trustworthiness and Safety of Multimodal Foundation Models

AutoRedTeamer: Autonomous Red Teaming with Lifelong Attack Integration

C-SafeGen: Certified Safe LLM Generation with Claim-Based Streaming Guardrails

Anyprefer: An Agentic Framework for Preference Data Synthesis

Fine-Grained Verifiers: Preference Modeling as Next-token Prediction in Vision-Language Alignment

ShieldAgent: Shielding Agents via Verifiable Safety Policy Reasoning

MJ-Bench: Is Your Multimodal Reward Model Really a Good Judge for Text-to-Image Generation?

Enhancing Vision-Language Model Reliability with Uncertainty-Guided Dropout Decoding

Efficient Multi-modal Large Language Models via Progressive Consistency Distillation

SafeVision: Efficient Image Guardrail with Robust Policy Adherence and Explainability

Preference Optimization with Multi-Sample Comparisons

Can Editing LLMs Inject Harm?

RANKCLIP: Ranking-Consistent Language-Image Pretraining