Search: “safety”

arXiv:2606.08531v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly evolving from simple text-based interaction systems into LLM agents that can maintain memory, use tools, a

arXiv:2606.05952v1 Announce Type: new Abstract: In this work, we propose an agentic gamification framework for hazard-informed learning of robot safety policies through synthetic scenarios. We model s

arXiv:2606.06423v1 Announce Type: new Abstract: Safety-critical traffic scenario generation is essential for evaluating autonomous driving systems under rare but high-risk interactions. Existing diffu

arXiv:2606.03954v1 Announce Type: new Abstract: As AI systems increasingly assist humans in physical tasks, ensuring safety becomes paramount -- physical actions carry immediate and irreversible conse

arXiv:2606.00651v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) architectures scale Large Language Models (LLMs) efficiently, enabling greater capacity with reduced computational cost by dy

arXiv:2512.20806v3 Announce Type: replace Abstract: Ensuring the safety of language models (LMs) while maintaining their usefulness remains a critical challenge in AI alignment. Current approaches rel

arXiv:2605.29396v1 Announce Type: new Abstract: Safety alignment for large language models (LLMs) aims to reduce harmful or unsafe behavior while preserving general utility. However, recent findings r

arXiv:2605.28830v1 Announce Type: cross Abstract: As Large Language Models (LLMs) are increasingly deployed in safety-critical applications, robust content moderation becomes essential. We present a c

arXiv:2605.28583v1 Announce Type: cross Abstract: Ensuring both safety and efficiency in decision-making for autonomous driving systems remains a fundamental challenge. Traditional Deep Reinforcement

arXiv:2605.26315v1 Announce Type: cross Abstract: Direct Preference Optimisation (DPO) is widely used for safety alignment in large language models. However, prior work shows it is brittle and exhibit

arXiv:1908.05972v3 Announce Type: replace Abstract: This paper significantly improves on, and finishes to validate, an approach proposed in previous research in which safety outcomes were predicted fr

arXiv:2605.21225v1 Announce Type: new Abstract: We address the problem of making a pre-trained reinforcement learning (RL) policy safety-aware by incorporating cost constraints without retraining it f

arXiv:2512.16856v2 Announce Type: replace Abstract: AI safety and alignment research has predominantly been focused on methods for safeguarding individual AI systems, resting on the assumption of an e

arXiv:2605.15782v1 Announce Type: new Abstract: In this work, we address the problem of ensuring real-time safety in autonomous robot navigation, in spatially constrained dynamic environments, by util

arXiv:2605.12386v1 Announce Type: new Abstract: Robotic manipulation is typically evaluated by task success, but successful completion does not guarantee safe execution. Many safety failures are tempo

arXiv:2601.18061v3 Announce Type: replace Abstract: Learning from human feedback~(LHF) assumes that expert judgments, appropriately aggregated, yield valid ground truth for training and evaluating AI

arXiv:2605.10386v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) are increasingly integrated into autonomous driving (AD) systems; however, they remain vulnerable to diverse sa

arXiv:2605.06992v1 Announce Type: new Abstract: AI agents are increasingly deployed in multi-task settings, where the task to perform is specified at test time, and the agent must generalize to unseen

arXiv:2605.04222v1 Announce Type: cross Abstract: Real-world control systems must achieve long-horizon objectives (liveness) while respecting continuous-time safety constraints, a combination that mot

arXiv:2603.10572v2 Announce Type: replace Abstract: Partially Observable Markov Decision Processes (POMDPs) provide a principled framework for robot decision-making under uncertainty. Solving reach-av

arXiv:2605.00227v1 Announce Type: new Abstract: There are growing concerns about the risks posed by AI companion applications designed for emotional engagement. Existing safety evaluations often rely

arXiv:2605.00326v1 Announce Type: new Abstract: Single-prompt first-token probabilities from zero-shot vision-language model (VLM) safety classifiers are treated as decision scores, but we show they a

arXiv:2604.27807v1 Announce Type: new Abstract: The design of embedded safety-critical systems such as those used in next-generation automotive and autonomous platforms, is increasingly challenged by

arXiv:2604.28087v1 Announce Type: cross Abstract: Rule-based systems remain central in safety-critical domains but often struggle with scalability, brittleness, and goal misspecification. These limita

arXiv:2505.10924v4 Announce Type: replace-cross Abstract: Recently, AI-driven interactions with computing devices have advanced from basic prototype tools to sophisticated, LLM-based systems that emul

arXiv:2604.26479v1 Announce Type: cross Abstract: Safety-critical prediction systems, such as autonomous vehicles, weather forecasters, and medical monitors, commonly rely on probabilistic forecasters

arXiv:2509.26238v4 Announce Type: replace Abstract: Monitoring large language models' (LLMs) activations is an effective way to detect harmful requests before they lead to unsafe outputs. However, tra

arXiv:2505.15404v2 Announce Type: replace Abstract: Large Reasoning Models (LRMs) have achieved remarkable success on reasoning-intensive tasks such as mathematics and programming. However, their enha

arXiv:2604.17691v1 Announce Type: new Abstract: Safety alignment in large language models is remarkably shallow: it is concentrated in the first few output tokens and reversible by fine-tuning on as f

arXiv:2604.16542v1 Announce Type: cross Abstract: Safety guardrails have become an active area of research in AI safety, aimed at ensuring the appropriate behavior of large language models (LLMs). How

arXiv:2604.14251v1 Announce Type: new Abstract: Monitoring LLM safety at scale requires balancing cost and accuracy: a cheap latent-space probe can screen every input, but hard cases should be escalat

arXiv:2604.12308v1 Announce Type: new Abstract: Individuals' concerns about data privacy and AI safety are highly contextualized and extend beyond sensitive patterns. Addressing these issues requires

arXiv:2604.12710v1 Announce Type: cross Abstract: Large language models (LLMs) often demonstrate strong safety performance in high-resource languages, yet exhibit severe vulnerabilities when queried i

arXiv:2507.01264v2 Announce Type: replace-cross Abstract: Designing diverse and safety-critical driving scenarios is essential for evaluating autonomous driving systems. In this paper, we propose a no

arXiv:2602.03402v3 Announce Type: replace Abstract: Vision language models (VLMs) extend the reasoning capabilities of large language models (LLMs) to cross-modal settings, yet remain highly vulnerabl

arXiv:2604.10429v1 Announce Type: new Abstract: This paper considers the problem of zero-shot safety guarantees for cascade dynamical systems. These are systems where a subset of the states (the inner

arXiv:2604.08881v1 Announce Type: new Abstract: In real-world deployments, Vision-Language Large Models (VLLMs) face critical challenges from multilingual and multimodal composite attacks: harmful ima

arXiv:2610.11483v1 Announce Type: new Abstract: Diffusion-based driving planners capture diverse behaviors but can generate unsafe trajectories under distribution shift. We propose BridgeGuard, a safe

arXiv:2610.09294v1 Announce Type: cross Abstract: Rapid progress in AI agents has brought growing attention to agent safety, with extensive evaluation focused on digital environments. As agents move i

arXiv:2610.07654v1 Announce Type: cross Abstract: On-policy distillation (OPD) has attracted growing attention as an effective way to transfer capabilities from teacher models to student models. Recen

arXiv:2603.04364v2 Announce Type: replace-cross Abstract: Multimodal web agents that process both screenshots and accessibility trees are increasingly deployed to interact with web interfaces, yet the

arXiv:2610.04084v1 Announce Type: new Abstract: Generative AI can produce engineering artefacts, but generation alone does not determine whether or how those artefacts should enter safety-critical wor

arXiv:2610.06306v1 Announce Type: new Abstract: General purpose language models are increasingly able to control robotic hardware. Understanding the capabilities and safety of these models when embodi

arXiv:2602.04448v3 Announce Type: replace Abstract: Mixture-of-Experts (MoE) language models introduce unique challenges for safety alignment due to their sparse routing mechanisms, which can enable d

arXiv:2610.00400v1 Announce Type: cross Abstract: Multi-turn attacks on agentic systems can compose individually permissible actions into harmful outcomes, challenging defenses that assess actions or

arXiv:2610.00093v1 Announce Type: cross Abstract: Large language models (LLMs) exhibit strong general capabilities, yet their parameters typically remain fixed after deployment, limiting learning from