Search: “safety”

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

14 Apr 2026

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

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

DGX agent

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

safetyarxiv-cs-ai

14 Apr 2026

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

14 Apr 2026

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

safetyarxiv-cs-cl

14 Apr 2026

Better Eyes, Better Thoughts: Why Vision Chain-of-Thought Fails in Medicine

arXiv:2603.06665v2 Announce Type: replace-cross Abstract: Large vision-language models (VLMs) often benefit from chain-of-thought (CoT) prompting in general domains, yet its efficacy in medical vision

CausalVAD: De-confounding End-to-End Autonomous Driving via Causal Intervention

arXiv:2603.18561v2 Announce Type: replace Abstract: Planning-oriented end-to-end driving models show great promise, yet they fundamentally learn statistical correlations instead of true causal relatio

ECHO: Efficient Chest X-ray Report Generation with One-step Block Diffusion

arXiv:2604.09450v1 Announce Type: cross Abstract: Chest X-ray report generation (CXR-RG) has the potential to substantially alleviate radiologists' workload. However, conventional autoregressive visio

Large Language Models Generate Harmful Content Using a Distinct, Unified Mechanism

arXiv:2604.09544v1 Announce Type: cross Abstract: Large language models (LLMs) undergo alignment training to avoid harmful behaviors, yet the resulting safeguards remain brittle: jailbreaks routinely

On the Spectral Geometry of Cross-Modal Representations: A Functional Map Diagnostic for Multimodal Alignment

arXiv:2604.08579v1 Announce Type: cross Abstract: We study cross-modal alignment between independently pretrained vision (DINOv2) and language (all-MiniLM-L6-v2) encoders using the functional map fram

Semantic Intent Fragmentation: A Single-Shot Compositional Attack on Multi-Agent AI Pipelines

arXiv:2604.08608v1 Announce Type: cross Abstract: We introduce Semantic Intent Fragmentation (SIF), an attack class against LLM orchestration systems where a single, legitimately phrased request cause

Are GUI Agents Focused Enough? Automated Distraction via Semantic-level UI Element Injection

arXiv:2604.07831v1 Announce Type: cross Abstract: Existing red-teaming studies on GUI agents have important limitations. Adversarial perturbations typically require white-box access, which is unavaila

safetyarxiv-cs-cl

Learning Without Losing Identity: Capability Evolution for Embodied Agents

arXiv:2604.07799v1 Announce Type: new Abstract: Embodied agents are expected to operate persistently in dynamic physical environments, continuously acquiring new capabilities over time. Existing appro

safetyarxiv-cs-ro

Machine Unlearning in the Era of Quantum Machine Learning: An Empirical Study

arXiv:2512.19253v4 Announce Type: replace-cross Abstract: We present the first empirical study of machine unlearning (MU) in hybrid quantum-classical neural networks. While MU has been extensively exp

SymptomWise: A Deterministic Reasoning Layer for Reliable and Efficient AI Systems

arXiv:2604.06375v1 Announce Type: new Abstract: AI-driven symptom analysis systems face persistent challenges in reliability, interpretability, and hallucination. End-to-end generative approaches ofte

AI Guardrail Survival under Single-Cycle Agentic Self-Summarization

arXiv:2608.11392v1 Announce Type: cross Abstract: Long-running agents periodically compact their context, replacing the transcript with a model-generated summary.Recent work shows that dropping a stan

Clustered Randomized Smoothing for Stochastic Prediction Functions

arXiv:2608.12037v1 Announce Type: new Abstract: Modern stochastic predictors can model rich, multi-modal outcome distributions. However, this expressive power comes with challenges in ensuring robust

safetyarxiv-cs-lg

Confidence Calibration of Deep Learning Systems

arXiv:2608.12100v1 Announce Type: cross Abstract: In high-stakes applications, reliable confidence estimates are as important as the predictions themselves. Confidence calibration ensures that predict

Constructing Dynamic Master Logic Models as Knowledge Graphs for Complex System Diagnostics Using Retrieval-Augmented Large Language Models

arXiv:2608.12304v1 Announce Type: new Abstract: Dynamic Master Logic (DML) provides a hierarchical framework for representing system behavior by linking functional objectives to underlying structural

Contextual Quality-Diversity Evolutionary Reinforcement Learning for HVAC Control in Tropical Commercial Buildings

arXiv:2608.11324v1 Announce Type: cross Abstract: This paper proposes a contextual quality-diversity evolutionary reinforcement-learning controller, CQD-ERL, for the supervisory control of a tropical,

Do Not Forget the Obvious - RISC: A Risk-Informed Slice-Coverage Protocol for Safe Autonomous Driving

arXiv:2608.12051v1 Announce Type: new Abstract: Aggregate metrics may not fully reflect performance in insufficiently examined high-risk driving conditions. We propose RISC (Risk-Informed Slice Covera

Faithful, Sufficient and Understandable: Rethinking Graph Counterfactual Explanations via Discrete Diffusion Inversion

arXiv:2608.12083v1 Announce Type: cross Abstract: Graph Neural Networks (GNNs) achieve strong predictive performance on graph-structured data across domains such as chemistry, biology, and network ana

Forecasting Side Effects of Activation Steering

arXiv:2608.11227v1 Announce Type: new Abstract: Activation steering modifies a language model by adding a learned direction to its hidden activations, enabling targeted behavioral changes without retr

GRPO for Financial Advice Generation: Outperforming Commercial LLMs under CATE Evaluation

arXiv:2608.11787v1 Announce Type: cross Abstract: Generating actionable financial advice from business records demands that models integrate numerical reasoning, domain knowledge, and sound judgment,

Is Per-Agent Policy Composition Safe? Rethinking Successor-Feature Transfer in Cooperative Multi-Agent Reinforcement Learning

arXiv:2608.11658v1 Announce Type: cross Abstract: Many reinforcement learning systems, from fleet management to traffic signal control, must serve an objective that changes dynamically after deploymen

Multi-Agent Embodied Autonomous Driving: From V2X Information Exchange to Shared World Models

arXiv:2606.13840v2 Announce Type: replace-cross Abstract: Autonomous driving is shifting from isolated vehicle intelligence toward multi-agent embodied systems that share perception, infer intent, and

On the Definition of Intelligence

arXiv:2507.22423v3 Announce Type: replace Abstract: To engineer AGI, we should first capture the essence of intelligence in a species-agnostic form that can be evaluated, while being sufficiently gene

TailBooster: A Dual-Layer Generative Framework for Extreme Value Augmentation with Operational Validity Enforcement

arXiv:2608.11951v1 Announce Type: cross Abstract: Extreme events in air transport, such as severe arrival delays and abnormal air times, cause cascading network disruptions with substantial operationa

Topology-Aware Query Selection for Surgical Instrument Instance Segmentation

arXiv:2608.11607v1 Announce Type: new Abstract: Accurate foreground masks can still form an incorrect surgical-instrument instance set: duplicate, fragmented, merged, missed, or empty-frame prediction

Unmasking Toxic Mimicry in Medical Offline Reinforcement Learning for ICU Sepsis Management via Counterfactual Clinical Audits

arXiv:2608.11410v1 Announce Type: new Abstract: Offline reinforcement learning (RL) offers considerable promise for optimizing ICU treatment decisions, yet standard evaluation metrics Mean Squared Err

safetyarxiv-cs-lg

A Convolutional Layer Activation Dimensionality Reduction for Out-of-Distribution and Adversarial Attack Detection Methods

arXiv:2608.10203v1 Announce Type: new Abstract: Despite the success of convolutional neural networks in image classification tasks and their general application in multi-modal models, their susceptibi

A Neural Network Based Teleoperation for Remote Controlled Vehicles

arXiv:2608.10367v1 Announce Type: new Abstract: Direct teleoperation of vehicles faces critical technical bottlenecks: communication latency and the operator's inability to physically perceive unmodel

safetyarxiv-cs-ro

ConfTriage: A Calibration-Aware LLM Triage Framework for Pulmonary Nodule Malignancy with Selective Specialist Deferral

arXiv:2608.10885v1 Announce Type: new Abstract: Pulmonary nodule malignancy prediction typically depends on image-trained specialist deep learning (DL) models that require substantial annotated imagin

Data Attribution of Emergent Misalignment with Persona Features

arXiv:2608.11025v1 Announce Type: new Abstract: Emergent misalignment (EM) is the phenomenon where fine-tuning a language model on a narrow task leads to harmful behavior in unrelated domains. A leadi

safetyarxiv-cs-cl

Dreamer-SAC: Off-Policy Learning in Latent World Models for Sample-Efficient Autonomous Driving

arXiv:2608.10386v1 Announce Type: new Abstract: Sample-efficient reinforcement learning for autonomous driving is often limited by the trade-off between data efficiency and model bias. While world mod

safetyarxiv-cs-lg

Local Ai

Easy3D-Labels: Supervising Semantic Occupancy Estimation with 3D Pseudo-Labels for Automotive Perception

arXiv:2509.26087v5 Announce Type: replace Abstract: In perception for automated vehicles, safety is critical not only for the driver but also for other agents in the scene, particularly vulnerable roa

local-aiarxiv-cs-cv

Expert-Guided g-computation with Large Language Models for Estimating Causal Effects on Timings: Applications to Hospital Quality Improvement

arXiv:2608.10339v1 Announce Type: cross Abstract: Hospital quality improvement (QI) programs routinely face multiple candidate interventions to optimize hospital flow, but existing methods struggle to

From Interpretability to Control: Insights from Six Years of the TrustNLP Workshop

arXiv:2608.11171v1 Announce Type: cross Abstract: The Workshop on Trustworthy Natural Language Processing (TrustNLP), co-located with major ACL conferences since 2021, has grown from 8 proceedings pap

LLMs Encode Their Failures: Predicting Success from Pre-Generation Activations

arXiv:2602.09924v4 Announce Type: replace-cross Abstract: Running LLMs with extended reasoning on every problem is expensive, but determining which inputs actually require additional compute remains c

Model Releases

REDAgentBench: Executable Red Teaming and Faithful Measurement of LLM Agent Systems

model-releasesarxiv-cs-ai

arXiv:2608.10669v1 Announce Type: new Abstract: Large language model (LLM) agents combine language-based reasoning with external tools to perform complex tasks. Adversarial inputs can exploit interact

Similarity Gates Approve Reversals: A Validity Audit of Embedding-Cosine Thresholds in Agent Systems

arXiv:2608.10216v1 Announce Type: cross Abstract: Agent frameworks ship quality gates that compare text blocks by embedding-cosine similarity and decide at a fixed cutoff. Deduplication filters, seman

Smart Enough to Go Extinct? An Evolutionary Challenge to the Value of General Intelligence and Its Ethical Implications for AGI

arXiv:2608.10730v1 Announce Type: cross Abstract: The pursuit of artificial general intelligence (AGI) rests on a seemingly self-evident premise: that general intelligence, the kind of flexible, domai

Toward a Theory of Value in AI Alignment

arXiv:2608.10327v1 Announce Type: new Abstract: Can AI systems be aligned to human values? The popularization of large language models (LLMs) and multi-modal foundation models has seen a rise in harms

UPAIR: Diagnosing Reasoning States via Uncertainty-Progress Alignment for Selective Intervention

arXiv:2607.17188v2 Announce Type: replace Abstract: While test-time scaling improves the problem-solving ability of large reasoning models (LRMs) through additional inference-time computation, it can

VIDS-Seg: Towards Reliable Uncertainty Quantification in Pediatric Cardiac Ultrasound Segmentation

arXiv:2608.10903v1 Announce Type: new Abstract: Reliable clinical deployment of machine learning requires models that know when they are likely to fail, particularly for subgroups underrepresented in

A Unified Framework for Dynamic Reward Shaping in Reinforcement Learning

arXiv:2608.08158v1 Announce Type: new Abstract: Sparse, delayed, and weakly informative rewards remain central obstacles to efficient reinforcement learning. Reward shaping addresses these limitations

Adaptive Sequential Test Planning for Multi-Mechanism Reliability Qualification via Bayesian Monte Carlo Tree Search

arXiv:2608.09622v1 Announce Type: new Abstract: Reliability qualification of advanced semiconductor devices requires sequential stress decisions that balance characterization objectives against multip

Agentic Harnesses: LLM-Driven Verification Layers for Robot Autonomy

arXiv:2608.09857v1 Announce Type: cross Abstract: Advances in advanced artificial intelligence tools have sparked research in robot autonomy, but the development of such systems has largely focused on

An Explainable GNN Framework for Component-Level Anomaly Diagnosis

arXiv:2608.09246v1 Announce Type: new Abstract: Industrial processes are complex systems composed of multiple interacting sensors that generate multivariate time series (MTS). Detecting anomalies in s