Search: “safety”

ODMA: On-Demand Memory Allocation Strategy for LLM Serving on LPDDR-Class Accelerators

arXiv:2512.09427v5 Announce Type: replace-cross Abstract: Existing memory management techniques severely hinder efficient Large Language Model serving on accelerators constrained by poor random-access

22 Apr 2026

Local Ai

Bridging the Culture Gap: A Framework for LLM-Driven Socio-Cultural Localization of Math Word Problems in Low-Resource Languages

DGX agent

arXiv:2508.14913v4 Announce Type: replace Abstract: Large language models (LLMs) have demonstrated significant capabilities in solving mathematical problems expressed in natural language. However, mul

local-aiarxiv-cs-cl

21 Apr 2026

Camo-M3FD: A New Benchmark Dataset for Cross-Spectral Camouflaged Pedestrian Detection

arXiv:2604.16582v1 Announce Type: new Abstract: Pedestrian detection is fundamental to autonomous driving, robotics, and surveillance. Despite progress in deep learning, reliable identification remain

Chain Of Interaction Benchmark (COIN): When Reasoning meets Embodied Interaction

model-releasesarxiv-cs-ro

arXiv:2604.16886v1 Announce Type: new Abstract: Generalist embodied agents must perform interactive, causally-dependent reasoning, continually interacting with the environment, acquiring information,

Emergent Misalignment via In-Context Learning: Narrow in-context examples can produce broadly misaligned LLMs

arXiv:2510.11288v4 Announce Type: replace Abstract: Recent work has shown that narrow finetuning can produce broadly misaligned LLMs, a phenomenon termed emergent misalignment (EM). While concerning,

ICAT: Incident-Case-Grounded Adaptive Testing for Physical-Risk Prediction in Embodied World Models

arXiv:2604.16405v1 Announce Type: cross Abstract: Video-generative world models are increasingly used as neural simulators for embodied planning and policy learning, yet their ability to predict physi

IncreFA: Breaking the Static Wall of Generative Model Attribution

arXiv:2604.17736v1 Announce Type: new Abstract: As AI generative models evolve at unprecedented speed, image attribution has become a moving target. New diffusion, adversarial and autoregressive gener

MedPRMBench: A Fine-grained Benchmark for Process Reward Models in Medical Reasoning

arXiv:2604.17282v1 Announce Type: new Abstract: Process-Level Reward Models (PRMs) are essential for guiding complex reasoning in large language models, yet existing PRM benchmarks cover only general

Privacy-R1: Privacy-Aware Multi-LLM Agent Collaboration via Reinforcement Learning

arXiv:2510.16054v2 Announce Type: replace-cross Abstract: When users submit queries to Large Language Models (LLMs), their prompts can often contain sensitive data, forcing a difficult choice: Send th

local-aiarxiv-cs-cl

Trustworthy Endoscopic Super-Resolution

arXiv:2604.18001v1 Announce Type: new Abstract: Super-resolution (SR) models are attracting growing interest for enhancing minimally invasive surgery and diagnostic videos under hardware constraints.

local-aiarxiv-cs-cv

Beyond Single-Model Optimization: Preserving Plasticity in Continual Reinforcement Learning

arXiv:2604.15414v1 Announce Type: cross Abstract: Continual reinforcement learning must balance retention with adaptation, yet many methods still rely on single-model preservation, committing to one e

20 Apr 2026

Power to the Clients: Federated Learning in a Dictatorship Setting

arXiv:2510.22149v3 Announce Type: replace-cross Abstract: Federated learning (FL) has emerged as a promising paradigm for decentralized model training, enabling multiple clients to collaboratively lea

20 Apr 2026

Syntheses

Wiki Lint Report — 2026-04-19

linthealth-checkautomated

Automated lint: 43 errors, 9 warnings, 3 info

19 Apr 2026

CausalDetox: Causal Head Selection and Intervention for Language Model Detoxification

arXiv:2604.14602v1 Announce Type: new Abstract: Large language models (LLMs) frequently generate toxic content, posing significant risks for safe deployment. Current mitigation strategies often degrad

Constrained Decoding for Safe Robot Navigation Foundation Models

model-releasesarxiv-cs-lg

arXiv:2509.01728v4 Announce Type: replace-cross Abstract: Recent advances in the development of robotic foundation models have led to promising end-to-end and general-purpose capabilities in robotic s

Dive into Claude Code: The Design Space of Today's and Future AI Agent Systems

arXiv:2604.14228v1 Announce Type: cross Abstract: Claude Code is an agentic coding tool that can run shell commands, edit files, and call external services on behalf of the user. This study describes

Rethinking Patient Education as Multi-turn Multi-modal Interaction

arXiv:2604.14656v1 Announce Type: cross Abstract: Most medical multimodal benchmarks focus on static tasks such as image question answering, report generation, and plain-language rewriting. Patient ed

SecureGate: Learning When to Reveal PII Safely via Token-Gated Dual-Adapters for Federated LLMs

arXiv:2602.13529v2 Announce Type: replace-cross Abstract: Federated learning (FL) enables collaborative training across organizational silos without sharing raw data, making it attractive for privacy-

local-aiarxiv-cs-cl

SelfGrader: Stable Jailbreak Detection for Large Language Models using Token-Level Logits

arXiv:2604.01473v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) are powerful tools for answering user queries, yet they remain highly vulnerable to jailbreak attacks. Existing g

Tight Sample Complexity Bounds for Best-Arm Identification Under Bounded Systematic Bias

arXiv:2604.14345v1 Announce Type: new Abstract: As search depth increases in autonomous reasoning and embodied planning, the candidate action space expands exponentially, heavily taxing computational

local-aiarxiv-cs-lg

AI Powered Image Analysis for Phishing Detection

arXiv:2604.13555v1 Announce Type: new Abstract: Phishing websites now rely heavily on visual imitation-copied logos, similar layouts, and matching colours-to avoid detection by text- and URL-based sys

Beyond Uniform Sampling: Synergistic Active Learning and Input Denoising for Robust Neural Operators

model-releasesarxiv-cs-lg

arXiv:2604.13316v1 Announce Type: new Abstract: Neural operators have emerged as fast surrogate models for physics simulations, yet they remain acutely vulnerable to adversarial perturbations, a criti

ROBOGATE: Adaptive Failure Discovery for Safe Robot Policy Deployment via Two-Stage Boundary-Focused Sampling

model-releasesarxiv-cs-ro

arXiv:2603.22126v3 Announce Type: replace Abstract: Deploying learned robot manipulation policies in industrial settings requires rigorous pre-deployment validation, yet exhaustive testing across high

Weakly-supervised Learning for Physics-informed Neural Motion Planning via Sparse Roadmap

arXiv:2604.13204v1 Announce Type: new Abstract: The motion planning problem requires finding a collision-free path between start and goal configurations in high-dimensional, cluttered spaces. Recent l

local-aiarxiv-cs-ro

CoD-Lite: Real-Time Diffusion-Based Generative Image Compression

arXiv:2604.12525v1 Announce Type: new Abstract: Recent advanced diffusion methods typically derive strong generative priors by scaling diffusion transformers. However, scaling fails to generalize when

Evaluating Language Models for Harmful Manipulation

arXiv:2603.25326v4 Announce Type: replace Abstract: Interest in the concept of AI-driven harmful manipulation is growing, yet current approaches to evaluating it are limited. This paper introduces a f

Generative Refinement Networks for Visual Synthesis

arXiv:2604.13030v1 Announce Type: new Abstract: While diffusion models dominate the field of visual generation, they are computationally inefficient, applying a uniform computational effort regardless

MoDora: Tree-Based Semi-Structured Document Analysis System

arXiv:2602.23061v3 Announce Type: replace-cross Abstract: Semi-structured documents integrate diverse interleaved data elements (e.g., tables, charts, hierarchical paragraphs) arranged in various and

Red Teaming Large Reasoning Models

arXiv:2512.00412v4 Announce Type: replace-cross Abstract: Large Reasoning Models (LRMs) have emerged as a powerful advancement in multi-step reasoning tasks, offering enhanced transparency and logical

SecureWebArena: A Holistic Security Evaluation Benchmark for LVLM-based Web Agents

arXiv:2510.10073v2 Announce Type: replace-cross Abstract: Large vision-language model (LVLM)-based web agents are emerging as powerful tools for automating complex online tasks. However, when deployed

Benchmarking Vision-Language Models under Contradictory Virtual Content Attacks in Augmented Reality

arXiv:2604.05510v2 Announce Type: replace Abstract: Augmented reality (AR) has rapidly expanded over the past decade. As AR becomes increasingly integrated into daily life, its security and reliabilit

Conflicts Make Large Reasoning Models Vulnerable to Attacks

arXiv:2604.09750v1 Announce Type: cross Abstract: Large Reasoning Models (LRMs) have achieved remarkable performance across diverse domains, yet their decision-making under conflicting objectives rema

From Translation to Superset: Benchmark-Driven Evolution of a Production AI Agent from Rust to Python

arXiv:2604.11518v1 Announce Type: cross Abstract: Cross-language migration of large software systems is a persistent engineering challenge, particularly when the source codebase evolves rapidly. We pr

GenTac: Generative Modeling and Forecasting of Soccer Tactics

arXiv:2604.11786v1 Announce Type: new Abstract: Modeling open-play soccer tactics is a formidable challenge due to the stochastic, multi-agent nature of the game. Existing computational approaches typ

How Alignment Routes: Localizing, Scaling, and Controlling Policy Circuits in Language Models

arXiv:2604.04385v3 Announce Type: replace-cross Abstract: This paper localizes the policy routing mechanism in alignment-trained language models. An intermediate-layer attention gate reads detected co

Pando: Do Interpretability Methods Work When Models Won't Explain Themselves?

arXiv:2604.11061v1 Announce Type: cross Abstract: Mechanistic interpretability is often motivated for alignment auditing, where a model's verbal explanations can be absent, incomplete, or misleading.

Hardware

Real-Time Voicemail Detection in Telephony Audio Using Temporal Speech Activity Features

arXiv:2604.09675v1 Announce Type: cross Abstract: Outbound AI calling systems must distinguish voicemail greetings from live human answers in real time to avoid wasted agent interactions and dropped c

hardwarearxiv-cs-ai

Retrieval-Augmented Large Language Models for Evidence-Informed Guidance on Cannabidiol Use in Older Adults

arXiv:2604.09548v1 Announce Type: cross Abstract: Older adults commonly experience chronic conditions such as pain and sleep disturbances and may consider cannabidiol for symptom management. Safe use

Aligned Agents, Biased Swarm: Measuring Bias Amplification in Multi-Agent Systems

arXiv:2604.08963v1 Announce Type: cross Abstract: While Multi-Agent Systems (MAS) are increasingly deployed for complex workflows, their emergent properties-particularly the accumulation of bias-remai

HiL-Bench (Human-in-Loop Benchmark): Do Agents Know When to Ask for Help?

arXiv:2604.09408v1 Announce Type: new Abstract: Frontier coding agents solve complex tasks when given complete context but collapse when specifications are incomplete or ambiguous. The bottleneck is n

Many-Tier Instruction Hierarchy in LLM Agents

arXiv:2604.09443v1 Announce Type: cross Abstract: Large language model agents receive instructions from many sources-system messages, user prompts, tool outputs, and more-each carrying different level

Neural networks for Text-to-Speech evaluation

arXiv:2604.08562v1 Announce Type: cross Abstract: Ensuring that Text-to-Speech (TTS) systems deliver human-perceived quality at scale is a central challenge for modern speech technologies. Human subje

Syntheses

Wiki Lint Report — 2026-04-12

linthealth-checkautomated

Automated lint: 34 errors, 0 warnings, 3 info

12 Apr 2026

Before We Trust Them: Decision-Making Failures in Navigation of Foundation Models

arXiv:2601.05529v5 Announce Type: replace Abstract: High success rates on navigation-related tasks do not necessarily translate into reliable decision making by foundation models. To examine this gap,

Beyond Surface Artifacts: Capturing Shared Latent Forgery Knowledge Across Modalities

arXiv:2604.07763v1 Announce Type: new Abstract: As generative artificial intelligence evolves, deepfake attacks have escalated from single-modality manipulations to complex, multimodal threats. Existi

CrashSight: A Phase-Aware, Infrastructure-Centric Video Benchmark for Traffic Crash Scene Understanding and Reasoning

arXiv:2604.08457v1 Announce Type: new Abstract: Cooperative autonomous driving requires traffic scene understanding from both vehicle and infrastructure perspectives. While vision-language models (VLM

Diagnosing and Mitigating Sycophancy and Skepticism in LLM Causal Judgment

arXiv:2601.08258v3 Announce Type: replace Abstract: Large language models increasingly fail in a way that scalar accuracy cannot diagnose: they produce a sound reasoning trace and then abandon it unde

Face-D(^2)CL: Multi-Domain Synergistic Representation with Dual Continual Learning for Facial DeepFake Detection

arXiv:2604.08159v1 Announce Type: new Abstract: The rapid advancement of facial forgery techniques poses severe threats to public trust and information security, making facial DeepFake detection a cri