Search: “safety”

Autonomous Topology Mutation: Safe Runtime Restructuring for Multi-Agent LLM Systems with Capability, State, and Shadow Invariants

arXiv:2607.20488v1 Announce Type: new Abstract: Multi-agent LLM frameworks typically fix their team topology at boot time. When an individual agent becomes overloaded at runtime, for example by mixing

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Local Ai

Concept Concentration for Faithful Representation Intervention

DGX agent

arXiv:2505.18672v2 Announce Type: replace Abstract: Representation intervention aims to localize and modify the representations that encode the underlying concepts in large language models (LLMs) to e

local-aiarxiv-cs-lg

24 Jul 2026

Local Ai

Conflict Resolution under Degraded Surveillance in Air Corridors Using Multi-Agent Reinforcement Learning

arXiv:2607.20547v1 Announce Type: new Abstract: Safe Advanced Air Mobility operations require aircraft to maintain separation when surveillance information is noisy, delayed, incomplete, or temporaril

local-aiarxiv-cs-lg

Geometric Configurations of Perturbed Jailbreak Prompts

arXiv:2607.20581v1 Announce Type: cross Abstract: Perturbation techniques that turn unsuccessful jailbreak prompts into successful ones are continuously evolving, constituting a major security threat

GuardianAgentBench: Where Agents Fail and How to Guard Them

arXiv:2607.20982v1 Announce Type: new Abstract: As large language model agents increasingly operate autonomously with access to tools and external environments, ensuring their safe and reliable behavi

ImplicitBBQ: Benchmarking Implicit Bias in Large Language Models through Characteristic Based Cues

arXiv:2604.01925v2 Announce Type: replace-cross Abstract: Large Language Models increasingly suppress biased outputs when demographic identity is stated explicitly, yet may still exhibit implicit bias

IssueTrojanBench: Benchmarking AI Coding Agents Against Malicious Issue Requests

arXiv:2607.20759v1 Announce Type: cross Abstract: AI coding agents powered by LLMs are increasingly integrated into real-world software development, where they generate, edit, and execute code with au

Refusal-Gated Decoding: Preserving Refusal Behavior Under High-Temperature Sampling

arXiv:2607.20791v1 Announce Type: new Abstract: High-temperature sampling is one of the primary mechanisms for increasing diversity in LLMs. Recent advances in truncation-based sampling techniques hav

Same Dangerous Objective, Opposite Advice: Direct Exposure versus Multi-Agent Mediation

arXiv:2607.21518v1 Announce Type: new Abstract: Even a current high-capability LLM can appear safer when shown a dangerous objective directly than when other agents transform and relay its direction.

SkillCorpus: Consolidating and Evaluating the Open Skill Ecosystem for Real-World LLM Agents

model-releasesarxiv-cs-cl

arXiv:2607.15557v4 Announce Type: replace Abstract: Agent skills, SKILL files that package reusable procedural knowledge for an LLM agent, are a popular mechanism for extending agent capabilities. Pub

StabilityBench: Benchmarking Instability in LLMs

arXiv:2607.20558v1 Announce Type: cross Abstract: AI Assistants are increasingly deployed in high-stakes settings, such as healthcare or government services. Yet their real-world behavior remains poor

WildTrace: Benchmarking Natural Evidence Trails in Long-Context Reasoning

arXiv:2607.09328v2 Announce Type: replace-cross Abstract: Answering complex questions over long documents frequently requires integrating evidence that the source itself disperses naturally across dis

ChannelGuard: Safe Models Do Not Compose into Safe Multi-Agent Systems

arXiv:2607.19430v1 Announce Type: cross Abstract: Multi-agent LLM applications chain a planner, worker agents, a verifier, and a synthesizer, and every hop between agents is an unmonitored channel thr

Emergent Autonomous Drifting for Collision Avoidance in Real-World Winter Driving Scenarios

arXiv:2607.19484v1 Announce Type: new Abstract: Real-world collision avoidance is a core motivation for studying the dynamics and control of high sideslip drifting in vehicles, yet the practical benef

FORCE-Bench: A Benchmark, Dataset, and Evaluation Harness for Agentic AI in Enterprise Finance

arXiv:2607.19409v1 Announce Type: new Abstract: Recent advances in large language models have accelerated deployment of agentic systems in operational finance. Existing benchmarks emphasize measuring

JailMeter: An Evidence-Based Evaluation Framework for Jailbreak Attacks on Large Language Models

arXiv:2607.19424v1 Announce Type: cross Abstract: The assessment of jailbreak attacks against large language models currently suffers from inconsistent evaluation criteria and methods, leading to unre

Train the Model, Not the Reader: Decodability Supervision for Verifiable Activation Explanations

arXiv:2607.20379v1 Announce Type: new Abstract: Natural-language autoencoders score explanations of hidden activations by reconstruction: an explanation is deemed faithful if the activation can be reg

Unified Prediction and Planning via Conflict-Aware Disjoint Parameter Training

arXiv:2607.19971v1 Announce Type: new Abstract: Accurate motion prediction of surrounding agents and safe motion planning are two closely coupled key tasks for social robot navigation in crowded envir

A Self-Evolving Agent for Longitudinal Personal Health Management

arXiv:2607.13940v1 Announce Type: new Abstract: Personal health management unfolds over repeated encounters, yet most health AI systems treat each request in isolation. We developed HealthClaw, an ope

CAVA: Canonical Action Verification and Attestation for Runtime Governance of Agentic AI Systems

arXiv:2607.13716v1 Announce Type: new Abstract: Agentic AI systems increasingly act through heterogeneous runtimes: local coding hooks, SDK tools, browser automation, managed-agent traces, API gateway

Efficient Text-to-Audio Generation via Pruning

arXiv:2607.13330v1 Announce Type: cross Abstract: Diffusion-based text-to-audio generative models such as AudioLDM achieve high perceptual quality and strong semantic consistency; however, their pract

HRIBench: Benchmarking Interaction-Centric Human-Robot Collaboration

arXiv:2607.13056v1 Announce Type: cross Abstract: Current vision-language-action (VLA) benchmarks primarily evaluate isolated manipulation skills while leaving human-robot interaction structure largel

MedRealMM: A Real-World Multimodal Benchmark for Chinese Online Medical Consultation

arXiv:2607.09142v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly deployed in online medical consultation, yet existing benchmarks remain poorly aligned with real clini

Not All Needles Are Found: How Fact Distribution and Don't Make It Up Prompts Shape Retrieval, Reasoning, and Hallucination in Long-Context LLMs

arXiv:2601.02023v2 Announce Type: replace-cross Abstract: As Large Language Models (LLMs) increasingly utilize massive context windows as working memory for autonomous tasks, their reliability fluctua

nuTruck: Benchmarking Autonomous Driving Planning for Distributed Electric-drive Trucks

arXiv:2607.13704v1 Announce Type: new Abstract: The dominance of traditional rule-based methods in autonomous driving has gradually been replaced by learning-based approaches. While learning-based pla

Securing LLMs in the Wild: Privacy and Security Challenges at the Edge

arXiv:2607.13088v1 Announce Type: cross Abstract: Large Language Models (LLMs) are rapidly moving from research settings into the wild, deployed on enterprise infrastructure, personal devices, and edg

Agents Don't Just Agree, They Remember: Benchmarking Persistent Sycophancy in Stateful Personal Agents

arXiv:2607.10526v2 Announce Type: replace Abstract: Stateful personal agents increasingly maintain long-term user profiles, episodic memories, and reusable skills. This persistence turns conversationa

Are we Merging the Right Models? Impact of Expert Training Duration on Model Merging for LLMs

arXiv:2607.11997v1 Announce Type: cross Abstract: Multi-task model merging combines separately trained expert models into a single model that handles all tasks without co-training. Standard practice m

Breaking Deja Vu: Independent Auditing of Visual Place Recognition through Vision-Language Reasoning

model-releasesarxiv-cs-cv

arXiv:2607.12818v1 Announce Type: new Abstract: Visual place recognition (VPR) is a key enabler of accurate localization and long-term autonomous navigation in robotics applications, such as loop clos

Continual Learning with Elastic Regularization and Synthetic Replay for Federated MLLM Fine-Tuning

arXiv:2607.12112v1 Announce Type: cross Abstract: Federated fine-tuning of Multimodal Large Language Models (MLLMs) across distributed networks enables privacy-sensitive adaptation to evolving data st

Epistemic Stance Flexibility Probing: Measuring Prompt-Conditioned Register Shift in Large Language Models

model-releasesarxiv-cs-cl

arXiv:2607.12739v1 Announce Type: new Abstract: A language model may be asked either what experts believe about a contested claim or what it believes about the claim itself. A trustworthy conversation

How Inference Compute Shapes Frontier LLM Evaluation

arXiv:2606.17930v2 Announce Type: replace Abstract: AI evaluations are shifting toward harder tasks that benefit from longer trajectories involving tool use and iterative problem solving. As a result,

How Many Tasks Are Enough for Agent Benchmark Decisions? A Replay Analysis of Public LLM Agent Benchmarks

arXiv:2607.12338v1 Announce Type: new Abstract: Agent benchmarks often compare two agents after all tasks have run, but costly evaluations make partial runs tempting. A task fraction alone does not sh

How to Realize Recursively Self-Improving Agents and Personal Singularity: A Goal-, Scope-, Tool-, and Benchmark-Driven Multi-Agent Architecture

model-releasesarxiv-cs-cv

arXiv:2607.12254v1 Announce Type: new Abstract: Large language model (LLM) agents can increasingly plan, use tools, maintain memory, and execute long-horizon tasks. These advances motivate two linked

Knowledge- and Gradient-Guided Reinforcement Learning for Parametrized Action Markov Decision Processes

arXiv:2607.12924v1 Announce Type: new Abstract: In this paper, we study Reinforcement Learning in Parametrized Action Markov Decision Processes (PAMDP), where each decision consists of a symbolic acti

Ontology-Amplified Distillation and Contextuality Auditing for Sovereign Enterprise Language Models: A Combined Proof-of-Mechanism and Negative-Results Method Study

arXiv:2607.11948v1 Announce Type: new Abstract: Regulated financial institutions operating under data-residency rules need tenant-owned language models that can run inside the institution's perimeter.

Local Ai

OmniOPSD: Rationale-Privileged On-Policy Self-Distillation for Affective Computing

arXiv:2606.15920v2 Announce Type: replace Abstract: Reinforcement learning for multimodal large language models (MLLMs) is often hindered by severe reward sparsity in complex reasoning tasks. This cha

local-aiarxiv-cs-cv

10 Jul 2026

Secure Decentralized Federated Learning via Gossip and Virtual Voting

arXiv:2607.08651v1 Announce Type: new Abstract: Decentralized federated learning (DFL) removes the central server by letting nodes exchange model updates through peer-to-peer gossip, but existing goss

10 Jul 2026

Shift & Drift: A Zero-Shot Benchmark for Generalizable and Robust Autonomous Driving Motion Planning

arXiv:2607.07844v1 Announce Type: cross Abstract: While closed-loop motion planners trained on large-scale, object-level datasets, e.g., nuPlan, demonstrate strong in-distribution (ID) performance, th

10 Jul 2026

Beyond Attack-Success Rate: Action-Graded Severity Scale for Tool-Using AI Agents

arXiv:2607.07474v1 Announce Type: cross Abstract: Agentic red-teaming benchmarks report whether an injected agent was compromised as a single bit: the attack succeeded, or it did not. We argue that th

Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages

arXiv:2607.06596v1 Announce Type: cross Abstract: Trusted monitoring is a central defense in AI control: a cheaper trusted model scores an untrusted model's actions for sabotage, and the most suspicio

Multi-Agent Robotic Control with Onboard Vision-Language Models

arXiv:2607.07403v1 Announce Type: cross Abstract: Vision Language Models (VLMs) and Vision Language Action (VLA) models have shown promise in robotic control. Yet, they face significant challenges reg

SmartHomeSecure: Automated Detection and Repair of Smart Home Configuration Errors Using Large Language Models

arXiv:2607.06748v1 Announce Type: cross Abstract: Smart home automation platforms increasingly rely on user-authored YAML configuration files to define device behaviors, but these files are prone to s

aiAuthZ: Off-Host, Identity-Bound Authorization for AI Agents

arXiv:2607.05518v1 Announce Type: cross Abstract: AI agents issue tool calls on the basis of text they cannot verify, so any party who controls part of the context can forge the appearance of authorit

8 Jul 2026

IMR: Iterative Mode-World Weighted Regression for Multi-Agent Trajectory Prediction

arXiv:2607.05705v1 Announce Type: cross Abstract: Multi-agent motion prediction is essential for automated vehicles to understand the intentions of surrounding vehicles. However, previous prediction-b

8 Jul 2026

A Reliable Context-Aware and Temporal Planning Framework for Autonomous Driving

model-releasesarxiv-cs-cv

arXiv:2607.04689v1 Announce Type: cross Abstract: Safe operation of autonomous vehicles in dense urban traffic depends on perception and planning that remain reliable when onboard sensing is degraded.

7 Jul 2026

A Technical Survey of Reinforcement Learning Techniques for Large Language Models

arXiv:2507.04136v2 Announce Type: replace Abstract: This survey offers a comprehensive foundation on the integration of RL with language models, highlighting prominent algorithms such as Proximal Poli

7 Jul 2026

Auditing the Audit: Five Failure Modes in Benchmark-Validity Audits

arXiv:2607.02586v1 Announce Type: new Abstract: Governance frameworks ask AI providers and auditors for documented evaluation evidence, and perturbation-based construct-validity audits are a common fo