7 papers
Evaluating and Explaining Prompt Sensitivity of LLMs Using Interactions
Ruiyang Qin, Qingzhuo Wang, Tian Wang +2
The remarkable capabilities of large language models (LLMs) are often undermined by their instability. Even subtle and semantically irrelevant changes in prompts can cause dramatic…
A Unified Approach to Interpreting Knowledge Distillation for Large Language Models via Interactions
Qingzhuo Wang, Ruiyang Qin, Zhenxin Qin +2
Despite the success of knowledge distillation (KD) in Large Language Models (LLMs), the underlying mechanism behind its efficacy remains unclear. In this paper, we propose a unifie…
HLL: Can Agents Cross Humanity's Last Line of Verification?
Xinhao Song, Su Su, Sirui Song +6
Multimodal agents are increasingly expected to operate interfaces on behalf of users, raising a central deployment question: can they truly substitute for humans in workflows that…
Multilingual Safety Alignment via Self-Distillation
Ruiyang Qin, Qingzhuo Wang, Dongrui Liu +3
Large language models (LLMs) exhibit severe multilingual safety misalignment: they possess strong safeguards in high-resource languages but remain highly vulnerable to jailbreak at…
Understanding and Defending VLM Jailbreaks via Jailbreak-Related Representation Shift
Zhihua Wei, Qiang Li, Jian Ruan +6
Large vision-language models (VLMs) often exhibit weakened safety alignment with the integration of the visual modality. Even when text prompts contain explicit harmful intent, add…
AgentDoG: A Diagnostic Guardrail Framework for AI Agent Safety and Security
Dongrui Liu, Qihan Ren, Chen Qian +40
The rise of AI agents introduces complex safety and security challenges arising from autonomous tool use and environmental interactions. Current guardrail models lack agentic risk…
FOLK: Fast Open-Vocabulary 3D Instance Segmentation via Label-guided Knowledge Distillation
Hongrui Wu, Zhicheng Gao, Jin Cao +3
Open-vocabulary 3D instance segmentation seeks to segment and classify instances beyond the annotated label space. Existing methods typically map 3D instances to 2D RGB-D images, a…