safety

Latent-GRPO: Group Relative Policy Optimization for Latent Reasoning

arXiv:2604.27998v1 Announce Type: cross Abstract: Latent reasoning offers a more efficient alternative to explicit reasoning by compressing intermediate reasoning into continuous representations and s

Learning from Disagreement: Clinician Overrides as Implicit Preference Signals for Clinical AI in Value-Based Care

DGX agent

arXiv:2604.28010v1 Announce Type: cross Abstract: We reframe clinician overrides of clinical AI recommendations as implicit preference data - the same signal structure exploited by reinforcement learn

safetyarxiv-cs-ai

1 May 2026

Learning Rate Transfer in Normalized Transformers

arXiv:2604.27077v1 Announce Type: cross Abstract: The Normalized Transformer, or nGPT (arXiv:2410.01131) achieves impressive training speedups and does not require weight decay or learning rate warmup

Learning Tactile-Aware Quadrupedal Loco-Manipulation Policies

arXiv:2604.27224v1 Announce Type: new Abstract: Quadrupedal loco-manipulation is commonly built on visual perception and proprioception. Yet reliable contact-rich manipulation remains difficult: visio

Learning-to-Explain through 20Q Gaming: An Explainable Recommender for Cybersecurity Education

arXiv:2604.26964v1 Announce Type: cross Abstract: The growing sophistication of contemporary cyber threats necessitates a more effective and adaptive approach to cybersecurity training. Intuitive and

Learning When to Remember: Risk-Sensitive Contextual Bandits for Abstention-Aware Memory Retrieval in LLM-Based Coding Agents

arXiv:2604.27283v1 Announce Type: cross Abstract: Large language model (LLM)-based coding agents increasingly rely on external memory to reuse prior debugging experience, repair traces, and repository

Linguistically Informed Multimodal Fusion for Vietnamese Scene-Text Image Captioning: Dataset, Graph Framework, and Phonological Attention

arXiv:2604.27712v1 Announce Type: cross Abstract: Scene-text image captioning requires fusing three information streams -- visual features, OCR-detected text, and linguistic knowledge -- to generate d

LLM Biases

arXiv:2604.26960v1 Announce Type: cross Abstract: Transformer-based agentic AI is rapidly being deployed on major platforms to help users shop, watch, and navigate content with less effort. While thes

Mapping how LLMs debate societal issues when shadowing human personality traits, sociodemographics and social media behavior

arXiv:2604.27624v1 Announce Type: cross Abstract: Large Language Models (LLMs) can strongly shape social discourse, yet datasets investigating how LLM outputs vary across controlled social and context

“Marcus’ specific point about coding is structurally important: a model that produces code which compiles and passes the tests it was given …

“Marcus’ specific point about coding is structurally important: a model that produces code which compiles and passes the tests it was given is not the same as a model that produces correct, secure, ma

Mechanized Foundations of Structural Governance: Machine-Checked Proofs for Governed Intelligence

arXiv:2604.27289v1 Announce Type: new Abstract: We present five results in the theory of structural governance for cognitive workflow systems. Three are mechanized in Coq 8.19 using the Interaction Tr

Meta is basically Black Mirror incarnate.

Meta is basically Black Mirror incarnate. This is a confusingly written piece, but the upshot is that Meta's smart glasses record even when you don't want them to, and that Meta's data analysis teams

METASYMBO: Multi-Agent Language-Guided Metamaterial Discovery via Symbolic Latent Evolution

arXiv:2604.27300v1 Announce Type: new Abstract: Metamaterial discovery seeks microstructured materials whose geometry induces targeted mechanical behavior. Existing inverse-design methods can efficien

MIFair: A Mutual-Information Framework for Intersectionality and Multiclass Fairness

arXiv:2604.28030v1 Announce Type: cross Abstract: Fairness in machine learning remains challenging due to its ethical complexity, the absence of a universal definition, and the need for context-specif

Mind the Gap: Structure-Aware Consistency in Preference Learning

arXiv:2604.27733v1 Announce Type: new Abstract: Preference learning has become the foundation of aligning Large Language Models (LLMs) with human intent. Popular methods, such as Direct Preference Opt

safetyarxiv-cs-lg

Mitigating Selection Bias in Large Language Models via Permutation-Aware GRPO

arXiv:2603.21016v2 Announce Type: replace-cross Abstract: Large language models (LLMs) used for multiple-choice and pairwise evaluation tasks often exhibit selection bias due to non-semantic factors l

MotuBrain: An Advanced World Action Model for Robot Control

arXiv:2604.27792v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models achieve strong semantic generalization but often lack fine-grained modeling of world dynamics. Recent work explores

MSR:Hybrid Field Modeling for CT-MRI Rigid-Deformable Registration of the Cervical Spine with an Annotated Dataset

arXiv:2604.27654v1 Announce Type: new Abstract: Accurate CT-MRI registration of the cervical spine is essential for preoperative planning because this region is anatomically complex,highly variable,an

OmniRobotHome: A Multi-Camera Platform for Real-Time Multiadic Human-Robot Interaction

arXiv:2604.28197v1 Announce Type: cross Abstract: Human-robot collaboration has been studied primarily in dyadic or sequential settings. However, real homes require multiadic collaboration, where mult

One of the things I hate the most about this site is the consistent lack of nuance. That’s why everything is an argument, and progress here …

Gary Marcus critiques social media platforms for lacking nuance in discourse, which he identifies as a root cause of persistent arguments and stalled progress on the site. The post reflects concerns a

Online semi-supervised perception: Real-time learning without explicit feedback

arXiv:2604.27562v1 Announce Type: new Abstract: This paper proposes an algorithm for real-time learning without explicit feedback. The algorithm combines the ideas of semi-supervised learning on graph

safetyarxiv-cs-lg

OpAgent: Operator Agent for Web Navigation

arXiv:2602.13559v2 Announce Type: replace Abstract: To fulfill user instructions, autonomous web agents must contend with the inherent complexity and volatile nature of real-world websites. Convention

OpenAI o1 System Card

arXiv:2412.16720v2 Announce Type: replace Abstract: The o1 model series is trained with large-scale reinforcement learning to reason using chain of thought. These advanced reasoning capabilities provi

PALCAS: A Priority-Aware Intelligent Lane Change Advisory System for Autonomous Vehicles using Federated Reinforcement Learning

arXiv:2604.27118v1 Announce Type: cross Abstract: We present a priority-aware intelligent lane change advisory system based on multi-agent federated reinforcement learning, namely PALCAS, for autonomo

Performance-Driven QUBO for Recommender Systems on Quantum Annealers

arXiv:2410.15272v3 Announce Type: replace-cross Abstract: Quantum annealers offer a promising hardware platform for solving combinatorial optimization problems, especially those formulated as Quadrati

Policy-Grounded Safety Evaluation of 20 Large Language Models

arXiv:2507.14719v2 Announce Type: replace Abstract: As large language models (LLMs) become increasingly integrated into real-world applications, scalable and rigorous safety evaluation is essential. T

Political Bias Audits of LLMs Capture Sycophancy to the Inferred Auditor

arXiv:2604.27633v1 Announce Type: new Abstract: Large language models (LLMs) are commonly evaluated for political bias based on their responses to fixed questionnaires, which typically place frontier

Preserving Temporal Dynamics in Time Series Generation

arXiv:2604.27182v1 Announce Type: cross Abstract: Time-series data augmentation plays a crucial role in regression-oriented forecasting tasks, where limited data restricts the performance of deep lear

Real-Time GPU-Accelerated Monte Carlo Evaluation of Safety-Critical AEB Systems Under Uncertainty

arXiv:2604.27193v1 Announce Type: new Abstract: Automatic Emergency Braking (AEB) systems represent a safety-critical national interest, with the National Highway Traffic Safety Administration (NHTSA)

Residual Gaussian Splatting for Ultra Sparse-View CBCT Reconstruction

arXiv:2604.27552v1 Announce Type: new Abstract: While 3D Gaussian splatting (3DGS) offers explicit and efficient scene representations for cone-beam computed tomography reconstruction, conventional ph

RHyVE: Competence-Aware Verification and Phase-Aware Deployment for LLM-Generated Reward Hypotheses

arXiv:2604.28056v1 Announce Type: new Abstract: Large language models (LLMs) make reward design in reinforcement learning substantially more scalable, but generated rewards are not automatically relia

Robot Learning from Human Videos: A Survey

arXiv:2604.27621v1 Announce Type: cross Abstract: A critical bottleneck hindering further advancement in embodied AI and robotics is the challenge of scaling robot data. To address this, the field of

Safe Bilevel Delegation (SBD): A Formal Framework for Runtime Delegation Safety in Multi-Agent Systems

arXiv:2604.27358v1 Announce Type: new Abstract: As large language model (LLM) agents are deployed in high-stakes environments, the question of how safely to delegate subtasks to specialized sub-agents

Sam Altman is a master at insincerity, Author @GaryMarcus claims. 'He is a master of projecting insincerity, a master at telling the room wh…

Sam Altman is a master at insincerity, Author @GaryMarcus claims. 'He is a master of projecting insincerity, a master at telling the room what it wants to hear and not always truthful'. “Altman, sitti

Sample-efficient evidence estimation of score based priors for model selection

arXiv:2602.20549v2 Announce Type: replace-cross Abstract: The choice of prior is central to solving ill-posed imaging inverse problems, making it essential to select one consistent with the measuremen

SCOOP: A pro-AI dark money group backed by a powerful super PAC funded by execs tied to Palantir and OpenAI, has been secretly paying influe…

SCOOP: A pro-AI dark money group backed by a powerful super PAC funded by execs tied to Palantir and OpenAI, has been secretly paying influencers to push pro-AI, anti-China propaganda on TikTok and IG

Stable Behavior, Limited Variation: Persona Validity in LLM Agents for Urban Sentiment Perception

arXiv:2604.28048v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly used as proxies for human perception in urban analysis, yet it remains unclear whether persona prompting p

Stating that women do not have penises is conservative. Stating that the scientific method is superior to ancestral tribal dances for seekin…

Stating that women do not have penises is conservative. Stating that the scientific method is superior to ancestral tribal dances for seeking truth is conservative. Supporting a rational immigration p

safetyelon-musk--x

Supercharging Agenda Setting Research: The ParlaCAP Dataset of 28 European Parliaments and a Scalable Multilingual LLM-Based Classification

arXiv:2602.16516v2 Announce Type: replace Abstract: This paper introduces ParlaCAP, a large-scale dataset for analyzing parliamentary agenda setting across Europe, and proposes a cost-effective method

Taxon: Hierarchical Tax Code Prediction with Semantically Aligned LLM Expert Guidance

arXiv:2601.08418v2 Announce Type: replace-cross Abstract: Tax code prediction is a crucial yet underexplored task in automating invoicing and compliance management for large-scale e-commerce platforms

Test Before You Deploy: Governing Updates in the LLM Supply Chain

arXiv:2604.27789v1 Announce Type: cross Abstract: Large Language Models (LLMs) are increasingly used as core dependencies in software systems. However, the hosted LLM services evolve continuously thro

Test-Time Distillation for Continual Model Adaptation

arXiv:2506.02671v3 Announce Type: replace Abstract: Deep neural networks often suffer performance degradation upon deployment due to distribution shifts. Continual Test-Time Adaptation (CTTA) aims to

The Effects of Visual Priming on Cooperative Behavior in Vision-Language Models

arXiv:2604.27953v1 Announce Type: new Abstract: As Vision-Language Models (VLMs) become increasingly integrated into decision-making systems, it is essential to understand how visual inputs influence

The Field of Safe Motion: Operationalizing Affordances in the Field of Safe Travel Using Reachability Analysis

arXiv:2604.27168v1 Announce Type: new Abstract: We present the Field of Safe Motion (FSM), a quantitative safety model for determining whether a driver maintains a collision-free escape route, or 'out

The Likelihood Ratio Wall: Structural Limits on Accurate Risk Assessment for Rare Violence

arXiv:2604.27282v1 Announce Type: cross Abstract: Pretrial risk assessment tools are used on over one million U.S. defendants each year, yet their use for predicting rare violent re-offense faces a ba

safetyarxiv-cs-lg

The Two Boundaries: Why Behavioral AI Governance Fails Structurally

arXiv:2604.27292v1 Announce Type: new Abstract: Every system that performs effects has two boundaries: what it can do (expressiveness) and what governance covers (governance). In nearly all deployed A

Tokenmaxxing is stupid. Change my mind?

Gary Marcus critiques the AI industry's focus on scaling model parameters and training data (tokenmaxxing) as an inefficient approach to advancing AI capabilities. He argues that simply increasing tok

TouchGuide: Inference-Time Steering of Visuomotor Policies via Touch Guidance

arXiv:2601.20239v4 Announce Type: replace Abstract: Fine-grained and contact-rich manipulation remain challenging for robots, largely due to the underutilization of tactile feedback. To address this,