Search: “safety”

Auditing Generative Audio Calls for Known-Task Audio-LLM Evaluation

arXiv:2608.27817v1 Announce Type: cross Abstract: Speech and audio LLMs are often evaluated by asking whether a waveform prompt beats an automatic speech recognition (ASR) transcript. For known closed

31 Aug 2026

Fast Weight Attention for Continual Learning

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

DGX agent

arXiv:2608.27763v1 Announce Type: cross Abstract: Recurrent fast-weight memories and selective state-space models compress an expanding context into a fixed-size recurrent state, making the state tran

safetyarxiv-cs-cl

31 Aug 2026

11/ METR and Redwood’s work is extremely valuable, but it assesses model behavior and alignment. That is only half the picture. We also need…

11/ METR and Redwood’s work is extremely valuable, but it assesses model behavior and alignment. That is only half the picture. We also need an independent account of how OpenAI’s containment failed.

safetygary-marcus--x

30 Aug 2026

AgentFold: Closed-Loop Agentic Search for Protein Folding Model Design

arXiv:2608.26747v1 Announce Type: new Abstract: Scientific LLM agents have shown promise in literature reasoning, tool use, and experiment planning, but it remains unclear whether they can autonomousl

AirLLM: Diffusion Policy-based Adaptive LoRA for Remote Fine-Tuning of LLM over the Air

arXiv:2507.11515v2 Announce Type: replace-cross Abstract: Operating Large Language Models (LLMs) on edge devices is increasingly challenged by limited communication bandwidth and strained computationa

Cross-Platform Generalisation Failure in Mental Health Natural Language Processing: A Five-Axis Fairness Audit of Transformer Models on Social Media

arXiv:2608.26138v1 Announce Type: new Abstract: We introduce the Cross-Platform Fairness Evaluation (CPFE) framework -- a five-axis audit protocol covering discriminative performance, calibration, sta

Evaluating AI Generated Summaries for Cancer Patients

arXiv:2608.26154v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly being integrated into digital health platforms to generate summaries of complex medical data. Although t

RedEvoAgent: Automatic Red-Teaming Agent with Experience-Driven Skill Evolution

arXiv:2608.27439v1 Announce Type: cross Abstract: LLM-based agents are increasingly deployed in product-level execution harnesses, where jailbreaks can trigger harmful tool use and persistent state ch

Self-Generated Text Recognition: Quality Heuristics, Cross-Task Transfer, and Downstream Bias in LLM Evaluation

arXiv:2608.26159v1 Announce Type: cross Abstract: Self-Generated Text Recognition (SGTR)--the ability of an LLM to identify its own outputs--poses risks to AI safeguards that rely on LLMs as evaluator

3DGAA: Realistic and Robust 3D Gaussian-based Adversarial Attack for Autonomous Driving

arXiv:2507.09993v4 Announce Type: replace Abstract: Camera-based perception in connected and autonomous vehicles remains exposed to physical adversarial attacks. Prior attacks often either optimize im

safetyarxiv-cs-cv

DESCENT: Directed Edge Scene Encoding for Airport Surface Movement Prediction

model-releasesarxiv-cs-ro

arXiv:2608.26002v1 Announce Type: new Abstract: Advanced automation is a key technology for enhancing the safety of ground operations amidst the increasing density of commercial air traffic. While mot

Enterprise AI's real risk isn't autonomous agents. It's the complexity between them.

Presented by Gravitee Agent complexity is the insidious shadow lurking inside enterprises right now that needs a light shone on it.That’s because enterprises don't deploy a single agent and watch it r

safetyventurebeat-ai

From Memorization to Absorption: Mixed-Policy RL for Continual Knowledge Injection

arXiv:2608.25243v1 Announce Type: new Abstract: Continual knowledge injection is essential for keeping large language models up-to-date in a fast-evolving world. Existing methods rely on supervised fi

LAC: Linear and Angular Compliance for Humanoid Whole-body Control

arXiv:2608.25405v1 Announce Type: new Abstract: Real-world humanoid tasks involve physical interaction with objects and humans, yet current controllers either reject external forces as disturbances or

safetyarxiv-cs-ro

RAEM: Robust Autonomous Exploration for Multi-Floor Environments with a Quadruped Robot

arXiv:2608.25366v1 Announce Type: new Abstract: In this paper, we propose RAEM, a robust autonomous exploration framework for quadruped robots operating in multi-floor environments. Most existing grou

safetyarxiv-cs-ro

When Stale Constraints Go Unchecked: Budgeted Verification Failures in Inherited Agent Memory

arXiv:2608.25553v1 Announce Type: cross Abstract: An agent that inherits a consolidated memory may inherit a constraint that was true when written and has since been withdrawn by a newer authoritative

ADVERSA: Measuring Multi-Turn Guardrail Degradation and Judge Reliability in Large Language Models

arXiv:2603.10068v2 Announce Type: replace-cross Abstract: Most adversarial evaluations of large language model (LLM) safety assess single prompts and report binary pass/fail outcomes, which fails to c

AI Finds A Way

arXiv:2608.23875v1 Announce Type: new Abstract: Artificial Intelligence (AI) algorithms frequently learn creative and unexpected solutions, surprising even expert researchers who develop and study the

(ext{DNN})^2: Doubly Non-Negative Relaxations for Deep Neural Networks

arXiv:2608.24743v1 Announce Type: new Abstract: Existing linear program (LP) and semidefinite program (SDP) relaxations for rectified linear unit (ReLU) neural network (NN) verification yield overly-c

safetyarxiv-cs-lg

Generating Biomedical Fact-Checking Reports with RL-Enhanced Agentic Search

arXiv:2608.23811v1 Announce Type: new Abstract: Automated fact-checking is essential for ensuring the reliability of public health information, yet the biomedical domain poses unique challenges. Valid

Illuminating the Three Dogmas of Reinforcement Learning under Evolutionary Light

arXiv:2507.11482v5 Announce Type: replace Abstract: Artificial learning systems are graduating from passive learners to increasingly autonomous agents, lending pragmatic urgency to the question of wha

Knowing When to Ask for Help: Bayesian Self-Escalation in Hierarchical LLM Agents

arXiv:2608.24087v1 Announce Type: cross Abstract: Current LLM agent systems decide delegation before reasoning begins (a router picks a model) or after a response is complete (a verifier scores it and

RoG-DAgger: Rollout-Guided Post-Training for End-to-End Driving

arXiv:2608.24525v1 Announce Type: new Abstract: Recent end-to-end driving systems demonstrate strong performance on closed-loop benchmarks, yet are still predominantly trained on fixed expert-collecte

safetyarxiv-cs-ro

GuardianBench: A Same-Scene Instruction-Contrastive Benchmark for Latent Contextual Risk in Embodied AI

arXiv:2608.21928v1 Announce Type: new Abstract: In embodied AI, safety risk can be latent: a benign instruction and a safe scene become hazardous only when composed. Prior work has advanced embodied s

Mission-Aligned Learning-Informed Control of Autonomous Systems: Formulation and Foundations

arXiv:2507.04356v3 Announce Type: replace-cross Abstract: Research, innovation and practical capital investment have been increasing rapidly toward the realization of autonomous physical agents. This

Robust Multi-Agent Reinforcement Learning for Small UAS Separation Assurance under GPS Degradation and Spoofing

arXiv:2603.28900v2 Announce Type: replace-cross Abstract: We address robust separation assurance for small Unmanned Aircraft Systems (sUAS) under GPS degradation and spoofing via Multi-Agent Reinforce

Tel Aviv and NYC-based Alice, which works to protect AI models from misuse and rogue behavior, raised 140M led by Apax Digital at a valuation 'close to 1B' (Marissa Newman/Bloomberg)

Marissa Newman / Bloomberg: Tel Aviv and NYC-based Alice, which works to protect AI models from misuse and rogue behavior, raised 140M led by Apax Digital at a valuation “close to 1B” — An Israeli sta

safetytechmeme

The Compaction Cliff in Long-Running AI Agent Memory

arXiv:2608.22752v1 Announce Type: new Abstract: A safety rule and an episodic log compete for the same tokens in an AI agent's context. When the budget overflows, both are summarized at the same rate;

CoAnchor: Robust Collaborative Perception under Spatio-Temporal Misalignment via Object-Level Anchors

arXiv:2608.21055v1 Announce Type: cross Abstract: Collaborative perception extends the sensing range of a single vehicle by fusing observations from nearby agents, which improves the robustness of aut

Evaluating Large Language Model Performance on International Maritime Dangerous Goods Code Compliance

arXiv:2608.21036v1 Announce Type: new Abstract: The transport of dangerous goods by sea is a high-consequence activity governed by the International Maritime Dangerous Goods (IMDG) Code, a complex reg

Mind the Style: Impact of Communication Style on Human-Chatbot Interaction

arXiv:2602.17850v2 Announce Type: replace-cross Abstract: Conversational agents increasingly mediate everyday digital interactions, yet the effects of their communication style on user experience and

Neural-Primitive: An Efficient End-to-end Local Planner with Primitive-based Imitation Learning for Autonomous Flight

arXiv:2608.20948v1 Announce Type: cross Abstract: Autonomous flight in unknown cluttered environments is hindered by the computation-quality-memory trilemma of onboard trajectory generation. In this p

A knowledge-guided agentic framework for mitigating patient-context ambiguity in health queries

arXiv:2608.19875v1 Announce Type: cross Abstract: Patients often submit short, underspecified queries to healthcare chatbots that lack the patient-specific information needed to determine an appropria

Bounded Sovereignty and the Control Tax: Pricing AI Oversight When the Deployer Does Not Own the Model

arXiv:2608.19216v1 Announce Type: new Abstract: AI control research asks how to deploy models safely even when they may be misaligned, but many control protocols assume that the deployer can instrumen

Multi-Agent Orchestration with the Common-Sense Reasoning Capabilities of LLMs for Autonomous Driving

arXiv:2608.20129v1 Announce Type: cross Abstract: Autonomous vehicles require robust perception and decision-making capabilities to operate in diverse and unseen scenarios. While reinforcement learnin

NEW from me in the Guardian -- I agree with the AI company employees who wrote, in their 'Pacing the Frontier' letter, that the government s…

NEW from me in the Guardian -- I agree with the AI company employees who wrote, in their 'Pacing the Frontier' letter, that the government should be stepping in. But companies could also be doing more

safetygary-marcus--x

Reading Between the Pixels: An Inscriptive Jailbreak Attack on Text-to-Image Models

arXiv:2604.05853v3 Announce Type: replace Abstract: Modern text-to-image (T2I) models can now render legible, paragraph-length text, enabling a fundamentally new class of misuse. We identify and forma

safetyarxiv-cs-cv

Winning at all costs

Hardmaru and Breaking911 posted a video (14 hours ago) of a humanoid robot training for the “Robot Olympics” in Beijing. In the footage, the robot accelerated too fast, failed to brake, collided with

safetydavid-ha--x

`From Prompt to Perturbation': An Adaptive Framework for Voice-Based Jailbreaks on Audio LLMs

arXiv:2502.00735v4 Announce Type: replace-cross Abstract: As large language models (LLMs) are increasingly integrated into audio-based applications, growing concerns have emerged regarding their vulne

20 Aug 2026

Going with the Flow(s): Distinct Clusters Target Individuals of Interest to Russia

Written by: Gabby Roncone, Wesley Shields Overview Google Threat Intelligence Group (GTIG) is tracking three distinct suspected Russian cyber espionage threat clusters abusing legitimate authenticatio

safetygoogle-cloud-ai

20 Aug 2026

Robust Risk Under Evolving Uncertainty: A Wasserstein Counterpart of the Entropic Value-at-Risk

arXiv:2608.19073v1 Announce Type: new Abstract: An agent still learning its environment should be cautious while ignorant and bold once confident. The entropic value-at-risk captures this through a ro

20 Aug 2026

Beyond the Trace: Coupling an Interpretable Reasoning-State Readout to Native MoE Routing

arXiv:2608.17638v1 Announce Type: new Abstract: What a reasoning model writes is only a partial record of the process that produces it. We introduce a two-level internal readout for mixture-of-experts

It is genuinely painful to me that almost nobody — not even doomers who should care deeply — seems to be paying attention to this critical p…

It is genuinely painful to me that almost nobody — not even doomers who should care deeply — seems to be paying attention to this critical point that might have truly profound consequences for humanit

safetygary-marcus--x

Multi-turn Conversational AI from Text to Multimodal Interaction: Data, Models, Evaluation, and Open Challenges

arXiv:2608.17605v1 Announce Type: cross Abstract: Conversational AI is moving beyond isolated text prompts toward sustained, multimodal interaction. In real conversations, users clarify goals, revise

Quantifying Risk Under Evolving Uncertainty: Belief-Dependent Robustness for Safe Sequential Decision Making

arXiv:2608.17574v1 Announce Type: new Abstract: How cautious should an agent be while it is still learning its environment? We propose RATTL (Risk-Adversarial Total-Reward Learning), which ties cautio

The Curious Case of Exploding DecPOMDPs: Containing the Fire through Policy Counting

arXiv:2608.17749v1 Announce Type: new Abstract: Decentralised partially observable Markov decision processes (DecPOMDPs) provide a general framework for modelling multi-agent decision making under unc

ETHOS: Towards a Modular Ethics Framework for Clinical Multi-Agent Systems

arXiv:2608.15424v1 Announce Type: cross Abstract: The rapid adoption of large language models has enabled the development of clinical multi-agent systems (MAS) capable of integrating multimodal patien

18 Aug 2026

MM-BEV: Enhancing Timeliness by Computing Where and When it Matters