Yingchun Wang

GhostEI-Bench: Do Mobile Agent Resilience to Environmental Injection in Dynamic On-Device Environments?

A$^2$RM: Adversarial-Augmented Reward Model

VRPRM: Process Reward Modeling via Visual Reasoning

A Rigorous Benchmark with Multidimensional Evaluation for Deep Research Agents: From Answers to Reports

SafeEvalAgent: Toward Agentic and Self-Evolving Safety Evaluation of LLMs

Probing the robustness of large language models safety to latent perturbations

FreezeVLA: Action-Freezing Attacks against Vision-Language-Action Models

JailBound: Jailbreaking Internal Safety Boundaries of Vision-Language Models

Reflection-Bench: Evaluating Epistemic Agency in Large Language Models