safety

Skill-SD: Skill-Conditioned Self-Distillation for Multi-turn LLM Agents

arXiv:2604.10674v1 Announce Type: cross Abstract: Reinforcement learning (RL) has been widely used to train LLM agents for multi-turn interactive tasks, but its sample efficiency is severely limited b

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

SLALOM: Simulation Lifecycle Analysis via Longitudinal Observation Metrics for Social Simulation

DGX agent

arXiv:2604.11466v1 Announce Type: cross Abstract: Large Language Model (LLM) agents offer a potentially-transformative path forward for generative social science but face a critical crisis of validity

safetyarxiv-cs-ai

14 Apr 2026

Some opportunities you just can't turn down, like when @GaryMarcus wants to come talk about AI at BugBash. Fewer than 20 tickets left, come …

Gary Marcus, a prominent AI researcher and critic known for his skepticism of current deep learning approaches, was invited to speak at BugBash, an event hosted by Antithesis. The post, shared by Anti

safetygary-marcus--x

Spatiotemporal-Aware Bit-Flip Injection on DNN-based Advanced Driver Assistance Systems (extended version)

arXiv:2604.03753v2 Announce Type: replace-cross Abstract: Modern advanced driver assistance systems (ADAS) rely on deep neural networks (DNNs) for perception and planning. Since DNNs' parameters resid

Speaking to No One: Ontological Dissonance and the Double Bind of Conversational AI

arXiv:2604.10833v1 Announce Type: cross Abstract: Recent reports indicate that sustained interaction with conversational artificial intelligence (AI) systems can, in a small subset of users, contribut

StaMo: Unsupervised Learning of Generalizable Robot Motion from Compact State Representation

arXiv:2510.05057v2 Announce Type: replace-cross Abstract: A fundamental challenge in embodied intelligence is developing expressive and compact state representations for efficient world modeling and d

Steered LLM Activations are Non-Surjective

arXiv:2604.09839v1 Announce Type: new Abstract: Activation steering is a popular white-box control technique that modifies model activations to elicit an abstract change in output behavior. It has als

Stop Fixating on Prompts: Reasoning Hijacking and Constraint Tightening for Red-Teaming LLM Agents

arXiv:2604.05549v2 Announce Type: replace Abstract: With the widespread application of LLM-based agents across various domains, their complexity has introduced new security threats. Existing red-team

Structural Consequences of Policy-Based Interventions on the Global Supply Chain Network

arXiv:2604.11479v1 Announce Type: new Abstract: As global political tensions rise and the anticipation of additional tariffs from the United States on international trade increases, the issues of econ

Structural Gating and Effect-aligned Lag-resolved Temporal Causal Discovery Framework with Application to Heat-Pollution Extremes

arXiv:2604.10371v1 Announce Type: new Abstract: This study proposes Structural Gating and Effect-aligned Discovery for Temporal Causal Discovery (SGED-TCD), a novel and general framework for lag-resol

Structured Causal Video Reasoning via Multi-Objective Alignment

arXiv:2604.04415v2 Announce Type: replace Abstract: Human understanding of video dynamics is typically grounded in a structured mental representation of entities, actions, and temporal relations, rath

SVD-Prune: Training-Free Token Pruning For Efficient Vision-Language Models

arXiv:2604.11530v1 Announce Type: cross Abstract: Vision-Language Models (VLM) have revolutionized multimodal learning by jointly processing visual and textual information. Yet, they face significant

Taking a Pulse on How Generative AI is Reshaping the Software Engineering Research Landscape

arXiv:2604.11184v1 Announce Type: cross Abstract: Context: Software engineering (SE) researchers increasingly study Generative AI (GenAI) while also incorporating it into their own research practices.

Task2vec Readiness: Diagnostics for Federated Learning from Pre-Training Embeddings

arXiv:2604.10849v1 Announce Type: cross Abstract: Federated learning (FL) performance is highly sensitive to heterogeneity across clients, yet practitioners lack reliable methods to anticipate how a f

TCSA-UDA: Text-Driven Cross-Semantic Alignment for Unsupervised Domain Adaptation in Medical Image Segmentation

arXiv:2511.05782v2 Announce Type: replace Abstract: Unsupervised domain adaptation for medical image segmentation remains a significant challenge due to substantial domain shifts across imaging modali

Teaching the Teacher: The Role of Teacher-Student Smoothness Alignment in Genetic Programming-based Symbolic Distillation

arXiv:2507.22767v3 Announce Type: replace-cross Abstract: Obtaining human-readable symbolic formulas via genetic programming-based symbolic distillation of a deep neural network trained on a target da

Tesla Insurance update With the latest version of Safety Score (v3.0), every mile you drive with FSD Supervised enabled will receive a score…

Tesla Insurance update With the latest version of Safety Score (v3.0), every mile you drive with FSD Supervised enabled will receive a score of 100. This allows you to maintain a higher average safety

safetyelon-musk--x

The Augmentation Trap: AI Productivity and the Cost of Cognitive Offloading

arXiv:2604.03501v2 Announce Type: replace-cross Abstract: Experimental evidence confirms that AI tools raise worker productivity, but also that sustained use can erode the expertise on which those gai

The Paradox of Professional Input: How Expert Collaboration with AI Systems Shapes Their Future Value

arXiv:2504.12654v1 Announce Type: cross Abstract: This perspective paper examines a fundamental paradox in the relationship between professional expertise and artificial intelligence: as domain expert

The Past Is Not Past: Memory-Enhanced Dynamic Reward Shaping

arXiv:2604.11297v1 Announce Type: cross Abstract: Despite the success of reinforcement learning for large language models, a common failure mode is reduced sampling diversity, where the policy repeate

The Poisoned Apple Effect: Strategic Manipulation of Mediated Markets via Technology Expansion of AI Agents

arXiv:2601.11496v2 Announce Type: replace-cross Abstract: The integration of AI agents into economic markets fundamentally alters the landscape of strategic interaction. We investigate the economic im

The Price of Ignorance: Information-Free Quotation for Data Retention in Machine Unlearning

arXiv:2604.11511v1 Announce Type: cross Abstract: When users exercise data deletion rights under the General Data Protection Regulation (GDPR) and similar regulations, mobile network operators face a

THOM: Generating Physically Plausible Hand-Object Meshes From Text

arXiv:2604.02736v3 Announce Type: replace Abstract: Generating photorealistic 3D hand-object interactions (HOIs) from text is important for applications like robotic grasping and AR/VR content creatio

Thought Branches: Interpreting LLM Reasoning Requires Resampling

arXiv:2510.27484v2 Announce Type: replace-cross Abstract: Most work interpreting reasoning models studies only a single chain-of-thought (CoT), yet these models define distributions over many possible

TInR: Exploring Tool-Internalized Reasoning in Large Language Models

arXiv:2604.10788v1 Announce Type: cross Abstract: Tool-Integrated Reasoning (TIR) has emerged as a promising direction by extending Large Language Models' (LLMs) capabilities with external tools durin

Towards Adaptive Open-Set Object Detection via Category-Level Collaboration Knowledge Mining

arXiv:2604.11195v1 Announce Type: cross Abstract: Existing object detectors often struggle to generalize across domains while adapting to emerging novel categories. Adaptive open-set object detection

Tracing the Thought of a Grandmaster-level Chess-Playing Transformer

arXiv:2604.10158v1 Announce Type: new Abstract: While modern transformer neural networks achieve grandmaster-level performance in chess and other reasoning tasks, their internal computation process re

Training-Free Cross-Lingual Dysarthria Severity Assessment via Phonological Subspace Analysis in Self-Supervised Speech Representations

arXiv:2604.10123v1 Announce Type: new Abstract: Dysarthric speech severity assessment typically requires trained clinicians or supervised models built from labelled pathological speech, limiting scala

Trajectory-based actuator identification via differentiable simulation

arXiv:2604.10351v1 Announce Type: new Abstract: Accurate actuation models are critical for bridging the gap between simulation and real robot behavior, yet obtaining high-fidelity actuator dynamics ty

safetyarxiv-cs-ro

Triviality Corrected Endogenous Reward

arXiv:2604.11522v1 Announce Type: new Abstract: Reinforcement learning for open-ended text generation is constrained by the lack of verifiable rewards, necessitating reliance on judge models that requ

UNIGEOCLIP: Unified Geospatial Contrastive Learning

arXiv:2604.11668v1 Announce Type: new Abstract: The growing availability of co-located geospatial data spanning aerial imagery, street-level views, elevation models, text, and geographic coordinates o

Utilizing and Calibrating Hindsight Process Rewards via Reinforcement with Mutual Information Self-Evaluation

arXiv:2604.11611v1 Announce Type: new Abstract: To overcome the sparse reward challenge in reinforcement learning (RL) for agents based on large language models (LLMs), we propose Mutual Information S

Variance-Aware Prior-Based Tree Policies for Monte Carlo Tree Search

arXiv:2512.21648v2 Announce Type: replace-cross Abstract: Monte Carlo Tree Search (MCTS) has profoundly influenced reinforcement learning (RL) by integrating planning and learning in tasks requiring l

Variational Latent Entropy Estimation Disentanglement: Controlled Attribute Leakage for Face Recognition

arXiv:2604.11250v1 Announce Type: new Abstract: Face recognition embeddings encode identity, but they also encode other factors such as gender and ethnicity. Depending on how these factors are used by

VeriTrans: Fine-Tuned LLM-Assisted NL-to-PL Translation via a Deterministic Neuro-Symbolic Pipeline

arXiv:2604.10341v1 Announce Type: new Abstract: extbf{VeriTrans} is a reliability-first ML system that compiles natural-language requirements into solver-ready logic with validator-gated reliability.

VideoStir: Understanding Long Videos via Spatio-Temporally Structured and Intent-Aware RAG

arXiv:2604.05418v2 Announce Type: replace-cross Abstract: Scaling multimodal large language models (MLLMs) to long videos is constrained by limited context windows. While retrieval-augmented generatio

ViserDex: Visual Sim-to-Real for Robust Dexterous In-hand Reorientation

arXiv:2604.11138v1 Announce Type: cross Abstract: In-hand object reorientation requires precise estimation of the object pose to handle complex task dynamics. While RGB sensing offers rich semantic cu

Visual Enhanced Depth Scaling for Multimodal Latent Reasoning

arXiv:2604.10500v1 Announce Type: new Abstract: Multimodal latent reasoning has emerged as a promising paradigm that replaces explicit Chain-of-Thought (CoT) decoding with implicit feature propagation

VLMaterial: Vision-Language Model-Based Camera-Radar Fusion for Physics-Grounded Material Identification

arXiv:2604.11671v1 Announce Type: cross Abstract: Accurate material recognition is a fundamental capability for intelligent perception systems to interact safely and effectively with the physical worl

safetyarxiv-cs-ro

Warm-Started Reinforcement Learning for Iterative 3D/2D Liver Registration

arXiv:2604.10245v1 Announce Type: new Abstract: Registration between preoperative CT and intraoperative laparoscopic video plays a crucial role in augmented reality (AR) guidance for minimally invasiv

WARPED: Wrist-Aligned Rendering for Robot Policy Learning from Egocentric Human Demonstrations

arXiv:2604.10809v1 Announce Type: new Abstract: Recent advancements in learning from human demonstration have shown promising results in addressing the scalability and high cost of data collection req

safetyarxiv-cs-ro

Weird Generalization is Weirdly Brittle

arXiv:2604.10022v1 Announce Type: new Abstract: Weird generalization is a phenomenon in which models fine-tuned on data from a narrow domain (e.g. insecure code) develop surprising traits that manifes

What Factors Affect LLMs and RLLMs in Financial Question Answering?

arXiv:2507.08339v4 Announce Type: replace Abstract: Recently, large language models (LLMs) and reasoning large language models (RLLMs) have gained considerable attention from many researchers. RLLMs e

What's In My Human Feedback? Learning Interpretable Descriptions of Preference Data

arXiv:2510.26202v2 Announce Type: replace-cross Abstract: Human feedback can alter language models in unpredictable and undesirable ways, as practitioners lack a clear understanding of what feedback d

When Can You Poison Rewards? A Tight Characterization of Reward Poisoning in Linear MDPs

arXiv:2604.10062v1 Announce Type: new Abstract: We study reward poisoning attacks in reinforcement learning (RL), where an adversary manipulates rewards within constrained budgets to force the target

When simulations look right but causal effects go wrong: Large language models as behavioral simulators

arXiv:2604.02458v2 Announce Type: replace-cross Abstract: Behavioral simulation is increasingly used to anticipate responses to interventions. Large language models (LLMs) enable researchers to specif

When Valid Signals Fail: Regime Boundaries Between LLM Features and RL Trading Policies

arXiv:2604.10996v1 Announce Type: cross Abstract: Can large language models (LLMs) generate continuous numerical features that improve reinforcement learning (RL) trading agents? We build a modular pi

Why why why is it so hard to understand that stupid systems can still be dangerous?

Why why why is it so hard to understand that stupid systems can still be dangerous? @Graffitinights1 For the zillionth time, dumb systems that are empowered can dangerous. like this monstrosity, as an

safetygary-marcus--x