12 papers
AgentSafeEvo: Adversarial Co-Evolution for LLM Agent Safety
Chenxu Wang, Xiaoyu Wen, Han Qi +4
Large language model (LLM) agents accomplish complex tasks through multi-turn interactions with external tools. However, malicious content embedded in user queries, tool descriptio…
Fair ASR: Re-Evaluating Black-Box Jailbreaks under Shared Target-Call Budgets
Zhida He, Xiaoyu Wen, Han Qi +5
Reliable jailbreak evaluation is essential for assessing LLM safety, but most existing studies rely solely on attack success rate (ASR) without accounting for its dependence on att…
JailbreakSkill: Scaling Automated Red-Teaming with Reusable and Ever-Evolving Skills
Xiaoyu Wen, Jiajia Li, Zhida He +11
Automated red-teaming has produced a growing collection of attack strategies, yet they typically remain scattered across prompts and workflows, making them difficult to systematica…
Harmony in Diversity: Multi-domain Contrastive Policy Optimization for Large Reasoning Models
Zongji Yu, Wenshui Luo, Yiliu Sun +5
Post-training via Reinforcement Learning (RL) has enabled Large Reasoning Models (LRMs) to achieve strong performance in individual domain. However, real-world applications increas…
Self-ReSET: Learning to Self-Recover from Unsafe Reasoning Trajectories
Dongcheng Zhang, Yi Zhang, Yuxin Chen +3
Large Reasoning Models possess remarkable capabilities for self-correction in general domain; however, they frequently struggle to recover from unsafe reasoning trajectories under…
Internalizing Safety Understanding in Large Reasoning Models via Verification
Yi Zhang, Yuxin Chen, Leheng Sheng +4
While explicit Chain-of-Thought (CoT) empowers large reasoning models (LRMs), it enables the generation of riskier final answers. Current alignment paradigms primarily rely on exte…
Not All Turns Matter: Credit Assignment for Multi-Turn Jailbreaking
Zhida He, Xiaoyu Wen, Han Qi +7
Deploying LLMs in multi-turn dialogues facilitates jailbreak attacks that distribute harmful intent across seemingly benign turns. Recent training-based multi-turn jailbreak method…
Dynamic Adversarial Reinforcement Learning for Robust Multimodal Large Language Models
Yicheng Bao, Xuhong Wang, Qiaosheng Zhang +3
Despite their impressive capabilities, Multimodal Large Language Models (MLLMs) exhibit perceptual fragility when confronted with visually complex scenes. This weakness stems from…
MAGIC: A Co-Evolving Attacker-Defender Adversarial Game for Robust LLM Safety
Xiaoyu Wen, Zhida He, Han Qi +7
Ensuring robust safety alignment is crucial for Large Language Models (LLMs), yet existing defenses often lag behind evolving adversarial attacks due to their \textbf{reliance on s…
Epistemic Traps: Rational Misalignment Driven by Model Misspecification
Xingcheng Xu, Jingjing Qu, Qiaosheng Zhang +4
The rapid deployment of Large Language Models and AI agents across critical societal and technical domains is hindered by persistent behavioral pathologies including sycophancy, ha…
KALE: Enhancing Knowledge Manipulation in Large Language Models via Knowledge-aware Learning
Qitan Lv, Tianyu Liu, Qiaosheng Zhang +2
Despite the impressive performance of large language models (LLMs) pretrained on vast knowledge corpora, advancing their knowledge manipulation-the ability to effectively recall, r…
UniJEPA: Enhancing Robot Policy via Unified Continuous and Discrete Representation Learning
Jianke Zhang, Yucheng Hu, Yanjiang Guo +5
Building generalist robot policies that can handle diverse tasks in open-ended environments is a central challenge in robotics. To leverage knowledge from large-scale pretraining,…