7 papers
ContraSafe-VL: A Contrastive Safety Preference Dataset for Vision-Language Models
Di Wu, Yanyan Zhao, Xin Lu +3
Vision-Language Models (VLMs) are increasingly deployed as general-purpose multimodal assistants, making multimodal safety alignment an important challenge. Safety depends on the i…
Safety Geometry Collapse in Multimodal LLMs and Adaptive Drift Correction
Jiahe Guo, Xiangran Guo, Jiaxuan Chen +6
Multimodal large language models (MLLMs) often fail to transfer safety capabilities learned in the text modality to semantically equivalent non-text inputs, revealing a persistent…
Learning to Learn from Multimodal Experience
Xingyu Sui, Weixiang Zhao, Yongxin Tang +4
Experience-driven learning has emerged as a promising paradigm for enabling agents to improve from interaction trajectories by accumulating and reusing past experience. However, ex…
Rethinking Experience Utilization in Self-Evolving Language Model Agents
Weixiang Zhao, Yingshuo Wang, Yichen Zhang +6
Self-evolving agents improve by accumulating and reusing experience from past interactions. Existing work has largely focused on how experience is constructed, represented, and upd…
On Safety Risks in Experience-Driven Self-Evolving Agents
Weixiang Zhao, Yichen Zhang, Yingshuo Wang +8
Experience-driven self-evolution has emerged as a promising paradigm for improving the autonomy of large language model agents, yet its reliance on self-curated experience introduc…
STAR-S: Improving Safety Alignment through Self-Taught Reasoning on Safety Rules
Di Wu, Yanyan Zhao, Xin Lu +2
Defending against jailbreak attacks is crucial for the safe deployment of Large Language Models (LLMs). Recent research has attempted to improve safety by training models to reason…
Self-Foveate: Enhancing Diversity and Difficulty of Synthesized Instructions from Unsupervised Text via Multi-Level Foveation
Mingzhe Li, Xin Lu, Yanyan Zhao
Synthesizing high-quality instruction data from unsupervised text is a promising paradigm for training large language models (LLMs), yet automated methods for this task still exhib…