Search: “safety”

Breaking the Curse of Repulsion: Remoteness-Aware Control of Negative Off-Policy Updates

arXiv:2602.10430v2 Announce Type: replace-cross Abstract: Off-policy policy optimization reuses historical behavior, including negative-advantage samples that suppress known failures. We show that rep

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

Calibrated Partial Resets: Preventing Policy Collapse in Continual Reinforcement Learning

DGX agent

arXiv:2607.24996v1 Announce Type: cross Abstract: Neural networks are hindered by accumulating dormant neurons and loss of expressivity throughout training, particularly in non-stationary data setting

safetyarxiv-cs-ai

29 Jul 2026

CAST: Game Solvers as Turn-Level Teachers for LLM Agents

arXiv:2607.25308v1 Announce Type: cross Abstract: Training large language models (LLMs) to act in long-horizon games is a promising step toward generalist decision-making, yet reinforcement learning w

CIFNet: An Analytic Neural Learning Framework for Efficient and Calibrated Class-Incremental Learning

arXiv:2509.11285v2 Announce Type: replace-cross Abstract: Class-Incremental Learning (CIL) in deep neural networks is conventionally framed as an iterative gradient-based optimization problem, incurri

Construction-Driven Injection: Linguistically-Grounded Edit-Based Code-Mixing Fingerprints for Large Language Models

arXiv:2607.25633v1 Announce Type: cross Abstract: Large language models (LLMs) are costly intellectual assets that remain exposed to unauthorized redistribution and commercial misuse. Injected fingerp

Context Assembly as the Controlled Variable: A Control-Theoretic View of Harness Policies for Frozen LLM Agents

arXiv:2607.25408v1 Announce Type: new Abstract: A growing body of 2026 work applies control theory to LLM agents: Lyapunov-certified stability for tool-mediated controllers (Prinos et al., 'Stable Age

ContractHIL-HLS: Contract-Aligned Multi-Agent Workflow with Hardware-in-the-Loop Feedback for HLS Design

arXiv:2607.25283v1 Announce Type: new Abstract: This paper presents ContractHIL-HLS, a contract-aligned multi-agent workflow for practical high-level synthesis (HLS) engineering. The workflow makes th

Contrastive Weak-to-strong Generalization

arXiv:2510.07884v3 Announce Type: replace-cross Abstract: Weak-to-strong generalization provides a promising paradigm for scaling large language models (LLMs) by training stronger models on samples fr

CoRT: Counterfactual Replay for Token-Level Rubric-Guided Policy Optimization

arXiv:2607.25659v1 Announce Type: new Abstract: Rubric-based reinforcement learning enriches language model training by evaluating model outputs against explicit criteria. Yet in GRPO-style pipelines,

Dataset Poisoning Attacks on Behavioral Cloning Policies

arXiv:2511.20992v2 Announce Type: replace Abstract: Behavior Cloning (BC) is a popular framework for training sequential decision policies from expert demonstrations via supervised learning. As these

safetyarxiv-cs-lg

DDSNet: Dual-domain Symmetry-aware Network for PCSEL Property Prediction

arXiv:2607.24785v1 Announce Type: cross Abstract: Efficient exploration of the photonic crystal (PhC) lattice design space is essential for developing photonic crystal surface-emitting lasers. While c

Decentralized Scalable Exploration via Emergent Adaptive Levy Walks on Minimal-Sensing Platforms

arXiv:2607.25195v1 Announce Type: new Abstract: Efficient autonomous exploration with palm-sized nano-UAVs remains challenging due to severe limitations in sensing, computation, and flight endurance.

DensFiLM: Density-Conditioned Video Saliency for Crowd Scenes

arXiv:2607.25465v1 Announce Type: new Abstract: Video saliency models typically apply a single fixation strategy across crowd scenes, despite systematic changes in attention with crowd density. Sparse

Do Models Fake Alignment Without Clear Consequences?

arXiv:2607.24758v1 Announce Type: new Abstract: Large language models are capable of recognizing evaluation contexts and altering their behavior to reflect evaluator expectations rather than typical d

Does Runtime Topology Context Improve LLM-Generated Kubernetes Security Patches?

arXiv:2607.25995v1 Announce Type: cross Abstract: Kubernetes is central to the cloud-native ecosystem, orchestrating containerised workloads. Recent work suggests that large language models (LLMs) can

Egocentric Station Holding of Robotic Fish in Unknown Turbulent Background Flow

arXiv:2607.24860v1 Announce Type: new Abstract: Approaching a target position and holding station in flowing water is a fundamental and critical capability for robotic fish operating in natural aquati

Evaluation of forced alignment of code-mixed speech: the case of Hindi-English

arXiv:2607.25581v1 Announce Type: new Abstract: Code-mixed speech poses unique challenges to forced alignment: expanded inventories, orthographic errors, and speaker variation. We evaluate forced alig

safetyarxiv-cs-cl

Face De-Identification: A Domain-Centric Survey from Capture to Processing

arXiv:2607.25926v1 Announce Type: cross Abstract: Face de-identification (De-ID) aims to remove or conceal personally identifiable facial features in images or videos to prevent identity recognition w

Faces of Fairness: Examining Bias in Facial Expression Recognition Datasets and Models

arXiv:2502.11049v3 Announce Type: replace Abstract: Automated Facial Expression Recognition (FER), involves two critical aspects: data and model design. Both significantly influence bias and fairness

Fairness Is Not Enough: Auditing Competence and Intersectional Bias in AI-powered Resume Screening

arXiv:2507.11548v3 Announce Type: replace-cross Abstract: The use of publicly available generative AI systems for resume evaluation is often justified by the assumption that these tools reduce bias re

Fine-Grained Food Image Understanding via Target-Aware Data Alignment

arXiv:2607.25794v1 Announce Type: new Abstract: Fine-grained food visual--semantic understanding requires models to capture subtle distinctions across ingredients, cooking methods, doneness, color, te

From Dyad to Triad: Eliciting XAI Requirements in Stroke Rehabilitation

arXiv:2607.25423v1 Announce Type: cross Abstract: Eliciting explainable AI (XAI) requirements from stroke survivors presents a methodological challenge with direct implications for the design of trust

@GaryMarcus OpenAI and Anthropic desperately need to raise prices. We seem to be faced with one of three choices for them: 1. Bailout 2. Ban…

Gary Marcus has argued that both OpenAI and Anthropic must increase their pricing, citing the rapid decline in token costs and competition from open‑source models. He presents a binary choice: either

safetygary-marcus--x

Group Equivariant Diffusion for Anomaly Detection in Computational Cytology

arXiv:2607.25503v1 Announce Type: new Abstract: Computational cytology on whole-slide images is challenging because malignant cells are rare, heterogeneous, and annotated slides are scarce. Anomaly de

Handy dandy AI crisis flowchart from @klonick

Gary Marcus shared a “handy dandy AI crisis flowchart” created by Kate Klonick (@Klonick) on Saturday, July 28. The tweet also notes that he had planned to write an article about Hugging Face and Open

safetygary-marcus--x

Model Releases

Harm is not Universal: Community-Specific Toxicity Detection is Urgently Needed

model-releasesarxiv-cs-ai

arXiv:2607.24898v1 Announce Type: cross Abstract: State-of-the-art toxicity detectors for text-to-image generation adopt a one-size-fits-all approach: a single universal model applying fixed safety gu

Interpretable GOHR Agents via Sparse Autoencoders

arXiv:2607.25132v1 Announce Type: new Abstract: A central challenge in interpreting learned decision-making systems is to determine whether their internal representations contain concepts that help ex

safetyarxiv-cs-lg

IRIS: Reusable Identity Representations from Frozen LLMs for Entity Alignment

arXiv:2607.25579v1 Announce Type: cross Abstract: Entity alignment (EA) identifies entities across knowledge graphs (KGs) that refer to the same real-world object. Conventional EA methods mainly explo

Keypoint-Guided Optimal Transport: Models, Algorithms, and Applications

arXiv:2303.13102v2 Announce Type: replace Abstract: Existing Optimal Transport (OT) methods mainly derive the optimal transport plan/matching under the criterion of transport cost/distance minimizatio

Large Language Model for Operations Research Formulation Selection in Multi-Warehouse Inventory Allocation

arXiv:2607.25956v1 Announce Type: new Abstract: Multi-warehouse inventory allocation is typically formulated as a mixed-integer programming (MIP) problem, yet no single formulation consistently matche

Learning from the Unseen: Offline Reinforcement Learning with Hidden Actions

arXiv:2607.25241v1 Announce Type: cross Abstract: Standard offline reinforcement learning (RL) algorithms typically assume that the actions in the dataset are observed without error. However, in many

safetyarxiv-cs-lg

LGFNet: A CTC-Guided Local-Global Fusion Framework for Single-Channel Sleep Staging

arXiv:2607.25197v1 Announce Type: new Abstract: Sleep staging remains challenging due to long-range temporal dependencies, ambiguous stage transitions-particularly in N1-and substantial distribution s

LLM as Forecasting Planner: Training-Free Text Conditioning for Time-Series Foundation Models

arXiv:2607.24892v1 Announce Type: cross Abstract: Text-conditioned time-series forecasting predicts a series from both its numerical history and natural-language context, allowing forecasts to account

Med-SegLens: Latent-Level Model Diffing for Interpretable Medical Image Segmentation

arXiv:2602.10508v2 Announce Type: replace Abstract: Modern segmentation models achieve strong predictive performance but remain largely opaque, limiting our ability to diagnose failures, understand da

Motion-Acceleration Calibration and Compensation in IMUs without External Equipment for Attitude Estimation Filters

arXiv:2607.25784v1 Announce Type: new Abstract: Attitude estimation based on inertial sensing requires measurements of local angular velocities and local gravity via gyroscopes and accelerometers. How

Motion Generation With Environmental Constraints

arXiv:2607.25053v1 Announce Type: new Abstract: Robot motion planning faces challenges in high-dimensional spaces and uncertain environments, often constrained by the need for collision-free motions.

Multi-Sensor Alignment for Weather Simulations

arXiv:2607.25612v1 Announce Type: new Abstract: Perception tasks for autonomous vehicles need to work satisfactorily in adverse weather conditions. Due to lack of real-world weather datasets, weather

NEXT: Reasoning-Driven Video Recommendation via a Vision-Language Model

arXiv:2607.24789v1 Announce Type: cross Abstract: We present NEXT (Next-interest EXploration Transformer), a reasoning-driven video recommendation framework that reasons over the video a user has just

ODYSSE: Episode-wise Policy Optimization for Personalized Agentic Reasoning

arXiv:2607.25369v1 Announce Type: new Abstract: Agentic systems have rapidly advanced in their ability to interact with real-world environments, leverage external tools, and provide services for users

On the Use of Synthetic Data for Threshold Calibration in Face Recognition: Performance and Security Implications for Border Control Systems

arXiv:2607.25990v1 Announce Type: new Abstract: The recently deployed Entry/Exit System (EES) introduces large-scale biometric verification into European border control, requiring face recognition sys

OPERA: Offline Policy-guided Expert Routing and Adaptation for Universal Biomedical Image Analysis

arXiv:2607.25108v1 Announce Type: cross Abstract: Biomedical image analysis spans diverse modalities and tasks, yet real-world deployment is hindered by severe distribution shifts across scanners, pro

P3: Probabilistic Policy Propagation for Stable VAE-Based Robot Learning

arXiv:2607.25541v1 Announce Type: new Abstract: Variational Autoencoders are widely used to encode high-dimensional and noisy observations in robotics. However, their stochastic latent creates a misma

Phase Structure in Rotary Attention: A Spectral Framework for Semantic Continuity and Execution-Boundary Governance

arXiv:2607.25507v1 Announce Type: new Abstract: Transformer language models are usually analyzed through vector geometry, yet ordered context and rotary position encoding introduce explicit phase stru

safetyarxiv-cs-cl

Pictura: Perspective-View Self-Play at Scale for Driving

arXiv:2607.26005v1 Announce Type: cross Abstract: Self-play in simulation produces robust driving policies at scale. Demonstrations of such behavior have been made using privileged vectorized observat

pimathbf{R}^2: Reactive Real-time Flow Policies

arXiv:2607.26055v1 Announce Type: cross Abstract: Generalist manipulation policies increasingly take the form of action-chunking flow policies built on large pretrained backbones. Such chunks run open

PLATO: Pointer Learner for Agent and Task Openness

arXiv:2607.25082v1 Announce Type: new Abstract: Open agent systems (OASYS) are increasingly prevalent in real-world domains where the sets of agents and tasks change unpredictably over time. Such open

Quotient Dynamics, Effective Curvature, and Implicit Bias in Positive Quadratic Networks

arXiv:2607.25624v1 Announce Type: new Abstract: Positive quadratic networks admit the low-rank representation f_U(x)=x^top UU^top x, where Uinmathbb{R}^{dtimes r} is identifiable only up to right orth

Ranked by Position: Order Sensitivity as an Exploitable Attack Surface in LLM Listwise Recommenders

arXiv:2607.24869v1 Announce Type: cross Abstract: Large language models (LLMs) used as listwise rerankers in recommendation systems suffer from position bias when serializing candidate sets into promp

safetyarxiv-cs-cl