相关论文: Efficient Refusal Ablation in LLM through Optimal …

Vision Language Models (VLMs) can produce unintended and harmful content when exposed to adversarial attacks, particularly because their vision capabilities create new vulnerabilities. Existing defenses, such as input preprocessing,…

Reinforcement learning (RL) can enable task-oriented dialogue systems to steer the conversation towards successful task completion. In an end-to-end setting, a response can be constructed in a word-level sequential decision making process…

Communication-avoiding algorithms for Linear Algebra have become increasingly popular, in particular for distributed memory architectures. In practice, these algorithms assume that the data is already distributed in a specific way, thus…

Large Language Models (LLMs) have achieved remarkable success across domains such as healthcare, education, and cybersecurity. However, this openness also introduces significant security risks, particularly through embedding space…

Large Vision-Language Models (LVLMs) have shown remarkable capabilities across a wide range of multimodal tasks. However, their integration of visual inputs introduces expanded attack surfaces, thereby exposing them to novel security…

Reinforcement Learning (RL) algorithms for safety alignment of Large Language Models (LLMs), such as Direct Preference Optimization (DPO), encounter the challenge of distribution shift. Current approaches typically address this issue…

Safety alignment is indispensable for Large Language Models (LLMs) to defend threats from malicious instructions. However, recent researches reveal safety-aligned LLMs prone to reject benign queries due to the exaggerated safety issue,…

Safety-aligned LLMs suffer from two failure modes: jailbreak (answering harmful inputs) and over-refusal (declining benign queries). Existing vector steering methods adjust the magnitude of answer vectors, but this creates a fundamental…

To circumvent the alignment of large language models (LLMs), current optimization-based adversarial attacks usually craft adversarial prompts by maximizing the likelihood of a so-called affirmative response. An affirmative response is a…

Achieving robust safety alignment in large language models (LLMs) while preserving their utility remains a fundamental challenge. Existing approaches often struggle to balance comprehensive safety with fine-grained controllability at the…

Large language models (LLMs) are typically aligned to be harmless to humans. Unfortunately, recent work has shown that such models are susceptible to automated jailbreak attacks that induce them to generate harmful content. More recent LLMs…

Inference-time scaling methods rely on Process Reward Models (PRMs), which are often poorly calibrated and overestimate success probabilities. We propose, to our knowledge, the first use of conditional optimal transport for calibrating…

Reinforcement Learning (RL) applications in real-world scenarios must prioritize safety and reliability, which impose strict constraints on agent behavior. Model-based RL leverages predictive world models for action planning and policy…

Iterative jailbreak methods that repeatedly rewrite and input prompts into large language models (LLMs) to induce harmful outputs -- using the model's previous responses to guide each new iteration -- have been found to be a highly…

Recent studies show that gradient-based universal image jailbreaks on vision-language models (VLMs) exhibit little or no cross-model transferability, casting doubt on the feasibility of transferable multimodal jailbreaks. We revisit this…

Multi-agent formation as well as obstacle avoidance is one of the most actively studied topics in the field of multi-agent systems. Although some classic controllers like model predictive control (MPC) and fuzzy control achieve a certain…

Autonomous agentic AI systems powered by vision-language models (VLMs) are rapidly advancing toward real-world deployment, yet their cross-modal reasoning capabilities introduce new attack surfaces for adversarial manipulation that exploit…

For Large Language Models (LLMs) to be reliably deployed, models must effectively know when not to answer: abstain. Reasoning models, in particular, have gained attention for impressive performance on complex tasks. However, reasoning…

Jailbreaking attacks can effectively induce unsafe behaviors in Large Language Models (LLMs); however, the transferability of these attacks across different models remains limited. This study aims to understand and enhance the…

Autonomous highway driving demands a critical balance between proactive, efficiency-seeking behavior and robust safety guarantees. This paper proposes Language Action-guided Reinforcement Learning (LA-RL) with Safety Guarantees, a novel…