7 papers
ContainmentBench: Trace-Based Evaluation of Post-Exposure Containment in Tool-Using LLM Agents
Wenhao Lan, Shan Li, Xinhua Lai +3
Tool-using large language model (LLM) agents read untrusted content, maintain memory, delegate tasks, and invoke tools with external side effects. Terminal attack-success or policy…
From Refusal Geometry to Safety Geometry: Harmfulness--Refusal Coupling under Dynamic Adversarial Fine-Tuning
Wenhao Lan, Shan Li, Xinhua Lai +3
Safety alignment requires language models to refuse harmful requests without losing the ability to answer benign ones. Existing robustness evaluations, however, do not reveal wheth…
Taming Preference Mode Collapse via Directional Decoupling Alignment in Diffusion Reinforcement Learning
Chubin Chen, Sujie Hu, Jiashu Zhu +8
Recent studies have demonstrated significant progress in aligning text-to-image diffusion models with human preference via Reinforcement Learning from Human Feedback. However, whil…
Latent Temporal Discrepancy as Motion Prior: A Loss-Weighting Strategy for Dynamic Fidelity in T2V
Meiqi Wu, Bingze Song, Ruimin Lin +5
Video generation models have achieved notable progress in static scenarios, yet their performance in motion video generation remains limited, with quality degrading under drastic d…
Artifact-Aware Evaluation for High-Quality Video Generation
Chen Zhu, Jiashu Zhu, Yanxun Li +6
With the rapid advancement of video generation techniques, evaluating and auditing generated videos has become increasingly crucial. Existing approaches typically offer coarse vide…
Taming Hallucinations: Boosting MLLMs' Video Understanding via Counterfactual Video Generation
Zhe Huang, Hao Wen, Aiming Hao +6
Multimodal Large Language Models (MLLMs) have made remarkable progress in video understanding. However, they suffer from a critical vulnerability: an over-reliance on language prio…