Search: “safety”

The human brain🧠 is incredibly efficient because it only activates the specific neurons needed for a thought. Modern LLMs naturally try to …

The human brain🧠 is incredibly efficient because it only activates the specific neurons needed for a thought. Modern LLMs naturally try to do this too (> 95% of neurons in feedforward layers stay sile

safetydavid-ha--x

8 May 2026

Safety

Anatomy of a failure: When, how, and why deep vision fails in scientific domains

DGX agent

arXiv:2605.04231v1 Announce Type: new Abstract: Mirroring its ubiquity in popular media and all human activities, the use of deep learning (DL) is rapidly growing in scientific imaging modalities. How

safetyarxiv-cs-cv

7 May 2026

Beyond Fixed Thresholds and Domain-Specific Benchmarks for Explainable Multi-Task Classification in Autonomous Vehicles

arXiv:2605.04299v1 Announce Type: new Abstract: Scene understanding is a vital part of autonomous driving systems, which requires the use of deep learning models. Deep learning methods are intrinsical

Encoding Predictability and Legibility for Style-Conditioned Diffusion Policy

arXiv:2603.16368v2 Announce Type: replace-cross Abstract: Striking a balance between efficiency and transparent motion is a core challenge in human-robot collaboration, as highly expressive movements

From Reach to Insert: Tactile-Augmented Precision Assembly under Sub-Millimeter Tolerances

arXiv:2605.04649v1 Announce Type: new Abstract: High-precision assembly frequently involves tight-tolerance insertions, where even slight pose errors can cause jamming or excessive interaction forces,

safetyarxiv-cs-ro

@GaryMarcus is on fire lately... follow him for #AI

Gary Marcus is an AI researcher and public intellectual who frequently shares commentary and insights about artificial intelligence developments on social media. His posts on X (formerly Twitter) cove

safetygary-marcus--x

How Does the Lagrangian Guide Safe Reinforcement Learning through Diffusion Models?

arXiv:2602.02924v2 Announce Type: replace Abstract: Diffusion policy sampling enables reinforcement learning (RL) to represent multimodal action distributions beyond suboptimal unimodal Gaussian polic

I'm really excited about this as a new tool in our interpretability tool kit

I'm really excited about this as a new tool in our interpretability tool kit In a new paper, we present NLAs, an unsupervised method for converting an LLM's internal state into human-readable text. I'

safetyjan-leike--x

InterFuserDVS: Event-Enhanced Sensor Fusion for Safe RL-Based Decision Making

arXiv:2605.04355v1 Announce Type: new Abstract: Autonomous driving systems rely heavily on robust sensor fusion to perceive complex envi- ronments. Traditional setups using RGB cameras and LiDAR often

It'll be 'impossible to slow down the ASI race' until it (very) suddenly isn't.

This post discusses the dynamics of artificial superintelligence (ASI) development as a competitive race, arguing that competitive pressures make it difficult to slow progress until a critical inflect

safetyconnor-leahy--x

Model Releases

Manifold of Failure: Behavioral Attraction Basins in Language Models

model-releasesarxiv-cs-lg

arXiv:2602.22291v3 Announce Type: replace Abstract: While prior work has focused on projecting adversarial examples back onto the manifold of natural data to restore safety, we argue that a comprehens

Misaligned by Reward: Socially Undesirable Preferences in LLMs

arXiv:2605.05003v1 Announce Type: new Abstract: Reward models are a key component of large language model alignment, serving as proxies for human preferences during training. However, existing evaluat

Predict-then-Diffuse: Adaptive Response Length for Compute-Budgeted Inference in Diffusion LLMs

arXiv:2605.04215v1 Announce Type: new Abstract: Diffusion-based Large Language Models (D-LLMs) represent a promising frontier in generative AI, offering fully parallel token generation that can lead t

Road Risk Monitor: A Deployable U.S. Road Incident Forecasting System with Live Weather and Road-Level Tiles

arXiv:2605.04242v1 Announce Type: new Abstract: Nationwide road-incident forecasting is a systems problem before it is a modeling problem. A usable service must connect historical incident archives, h

Sparse Tokens Suffice: Jailbreaking Audio Language Models via Token-Aware Gradient Optimization

arXiv:2605.04700v1 Announce Type: cross Abstract: Jailbreak attacks on audio language models (ALMs) optimize audio perturbations to elicit unsafe generations, and they typically update the entire wave

Terence Tao recognized that plausibility and veracity are not the same, and that current tools are better at the former than the latter. [ed…

Terence Tao recognized that plausibility and veracity are not the same, and that current tools are better at the former than the latter. [edit: the video is from 2024, and affirms what i said in 2019

safetygary-marcus--x

A Knowledge-Driven LLM-Based Decision-Support System for Explainable Defect Analysis and Mitigation Guidance in Laser Powder Bed Fusion

arXiv:2605.01100v1 Announce Type: new Abstract: This work presents a knowledge-driven decision-support system that integrates structured defect knowledge with LLM-based reasoning to provide explainabl

Algebraic Semantics of Governed Execution: Monoidal Categories, Effect Algebras, and Coterminous Boundaries

arXiv:2605.01032v2 Announce Type: new Abstract: We present an algebraic semantics for governed execution in which governance is axiomatized, compositional, and coterminous with expressibility. The fra

Architectural Obsolescence of Unhardened Agentic-AI Runtimes

arXiv:2605.01740v1 Announce Type: cross Abstract: An agentic-AI runtime issues tool calls, sends messages, and actuates devices on behalf of an LLM. Catching the four ways an action can diverge from i

Audio-Visual Intelligence in Large Foundation Models

arXiv:2605.04045v1 Announce Type: new Abstract: Audio-Visual Intelligence (AVI) has emerged as a central frontier in artificial intelligence, bridging auditory and visual modalities to enable machines

Governing What the EU AI Act Excludes: Accountability for Autonomous AI Agents in Smart City Critical Infrastructure

arXiv:2605.01091v1 Announce Type: cross Abstract: When a traffic signal controller adjusts green phases and a grid manager curtails power on the same corridor, each system may comply with its own obli

Height Control and Optimal Torque Planning for Jumping With Wheeled-Bipedal Robots

arXiv:2605.03302v1 Announce Type: new Abstract: This paper mainly studies the accurate height jumping control of wheeled-bipedal robots based on torque planning and energy consumption optimization. Du

safetyarxiv-cs-ro

Logic-Constrained Shortest Paths for Flight Planning

arXiv:2412.13235v4 Announce Type: replace Abstract: The logic-constrained shortest path problem (LCSPP) combines a one-to-one shortest path problem with satisfiability constraints imposed on the routi

MILD: Mediator Agent System with Bidirectional Perception and Multi-Layered Alignment for Human-Vehicle Collaboration

arXiv:2605.01507v1 Announce Type: new Abstract: Prior studies report that partial driving automation can increase the cognitive demands on human drivers. This effect largely arises from human drivers'

Model Spec Midtraining: Improving How Alignment Training Generalizes

arXiv:2605.02087v1 Announce Type: new Abstract: Some frontier AI developers aim to align language models to a Model Spec or Constitution that describes the intended model behavior. However, standard a

NORA: A Harness-Engineered Autonomous Research Agent for End-to-End Spatial Data Science

arXiv:2605.02092v1 Announce Type: new Abstract: The automation of scientific research workflows has emerged as a transformative frontier in artificial intelligence, yet existing autonomous research ag

The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence

arXiv:2408.12622v3 Announce Type: replace-cross Abstract: Artificial intelligence (AI) is reshaping society, from video generation to medical diagnosis, coding agents to autonomous vehicles. Yet resea

Viewpoint-Agnostic Grasp Pipeline using VLM and Partial Observations

arXiv:2603.07866v2 Announce Type: replace-cross Abstract: Robust grasping in cluttered, unstructured environments remains challenging for mobile legged manipulators due to occlusions that lead to part

A Deep Learning Model for Battery State Prediction towards Intelligent Energy Management

arXiv:2605.00898v1 Announce Type: cross Abstract: Accurate forecasting of battery health indicators, including remaining capacity and lifetime, is of paramount importance for ensuring the reliability,

AgentReputation: A Decentralized Agentic AI Reputation Framework

arXiv:2605.00073v1 Announce Type: new Abstract: Decentralized, agentic AI marketplaces are rapidly emerging to support software engineering tasks such as debugging, patch generation, and security audi

Artificial intelligence language technologies in multilingual healthcare: Grand challenges ahead

arXiv:2605.01441v1 Announce Type: new Abstract: AI language technologies (AILTs), increasingly enabled by large language models (LLMs), are becoming embedded in multilingual healthcare workflows for t

Beyond Semantic Relevance: Counterfactual Risk Minimization for Robust Retrieval-Augmented Generation

arXiv:2605.01302v1 Announce Type: new Abstract: Standard Retrieval-Augmented Generation (RAG) systems predominantly rely on semantic relevance as a proxy for utility. However, this assumption collapse

Experience Constrained Hierarchical Federated Reinforcement Learning for Large-scale UAV Teams in Hazardous Environments

arXiv:2605.02165v1 Announce Type: new Abstract: Conventional federated learning assumes that greater learner participation improves training performance, by leveraging abundant, independently generate

Hazard-Aware Traffic Scene Graph Generation

arXiv:2603.03584v2 Announce Type: replace Abstract: Maintaining situational awareness in complex driving scenarios is challenging. It requires continuously prioritizing attention among extensive scene

Logit-Gap Steering: A Forward-Pass Diagnostic for Alignment Robustness

arXiv:2506.24056v2 Announce Type: replace-cross Abstract: RLHF-style alignment trains language models to refuse unsafe requests, but how much operational margin does this refusal rest on? We introduce

Machine Learning Enhanced Laser Spectroscopy for Multi-Species Gas Detection in Complex and Harsh Environments

arXiv:2605.01306v1 Announce Type: cross Abstract: Laser absorption spectroscopy (LAS) is a well-established technique for non-intrusive measurement of gas species in combustion and atmospheric environ

Patient-Specific Optimization for Mandibular Reconstruction Planning with Enhanced Bone Union

arXiv:2605.01084v1 Announce Type: new Abstract: Mandibular reconstruction with vascularized bone grafts is complicated by donor-host nonunion, and current virtual surgical planning produces a geometri

PRCD-MAP: Learning How Much to Trust Imperfect Priors in Causal Discovery

arXiv:2605.01669v1 Announce Type: cross Abstract: External priors of unknown reliability create a brittle trade-off in causal discovery: blind trust amplifies errors, blind rejection wastes signal. Re

Protein-Conditioned Multi-Objective Reinforcement Learning for Full-Length mRNA Design

arXiv:2605.01513v1 Announce Type: new Abstract: Designing therapeutic messenger RNA (mRNA) requires creating full-length transcripts that carefully balance stability, translation efficiency, and immun

Reinforcement Learning from Compiler and Language Server Feedback

arXiv:2510.22907v2 Announce Type: replace Abstract: Coding agents fail when text-level guesses outrun program facts: they hallucinate APIs, drift to the wrong symbol, and apply edits without evidence

Semantic Risk-Aware Heuristic Planning for Robotic Navigation in Dynamic Environments: An LLM-Inspired Approach

arXiv:2605.02862v1 Announce Type: new Abstract: The integration of Large Language Model (LLM) reasoning principles into classical robot path planning represents a rapidly emerging research direction.

safetyarxiv-cs-ro

SurgTEMP: Temporal-Aware Surgical Video Question Answering with Text-guided Visual Memory for Laparoscopic Cholecystectomy

arXiv:2603.29962v3 Announce Type: replace Abstract: Surgical procedures are inherently complex and risky, requiring extensive expertise and constant focus to navigate evolving intraoperative scenes. C

The crux of the Musk-OpenAI trial so far is that Brockman keeps acting as if selling chatbots and API access for profit (which is what they …

The crux of the Musk-OpenAI trial so far is that Brockman keeps acting as if selling chatbots and API access for profit (which is what they do now) is exactly the same as the original mission (which w

safetygary-marcus--x

Thermal Imaging for Contactless Cardiorespiratory and Sudomotor Response Monitoring

arXiv:2602.12361v2 Announce Type: replace Abstract: Human-machine interfaces in industrial automation need sensing modules that monitor operator actions and physiological state. This is important in f

Training-Free Adaptive 360-degree Video Streaming via Semantic Potential Fields

arXiv:2603.20999v2 Announce Type: replace-cross Abstract: Adaptive 360{eg} video streaming for teleoperation faces two coupled challenges: viewport prediction under uncertain gaze patterns and bitrate

Virtual Speech Therapist: A Clinician-in-the-Loop AI Speech Therapy Agent for Personalized and Supervised Therapy

arXiv:2605.01101v1 Announce Type: cross Abstract: This paper develops Virtual Speech Therapist (VST), an intelligent agent-based platform that streamlines stuttering assessment and delivers customized

Conformalized Quantum DeepONet Ensembles for Scalable Operator Learning with Distribution-Free Uncertainty

arXiv:2605.00330v1 Announce Type: new Abstract: Operator learning enables fast surrogate modeling of high-dimensional dynamical systems, but existing approaches face two fundamental limitations: quadr

4 May 2026

Exploring LLM biases to manipulate AI search overview

arXiv:2605.00012v1 Announce Type: cross Abstract: Modern large language models (LLMs) are used in many business applications in general, and specifically in web search systems and applications that ge