safety

NEO: No-Optimization Test-Time Adaptation through Latent Re-Centering

arXiv:2510.05635v2 Announce Type: replace-cross Abstract: Test-Time Adaptation (TTA) methods are often computationally expensive, require a large amount of data for effective adaptation, or are brittl

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

Neural at ArchEHR-QA 2026: One Method Fits All: Unified Prompt Optimization for Clinical QA over EHRs

DGX agent

arXiv:2605.10877v1 Announce Type: new Abstract: Automated question answering (QA) over electronic health records (EHRs) demands precise evidence retrieval, faithful answer generation, and explicit gro

safetyarxiv-cs-cl

12 May 2026

Neural Co-state Policies: Structuring Hidden States in Recurrent Reinforcement Learning

arXiv:2605.05373v2 Announce Type: replace Abstract: A key capability of intelligent agents is operating under partial observability: reasoning and acting effectively despite missing or incomplete stat

Neuromorphic Reinforcement Learning for Quadruped Locomotion Control on Uneven Terrain

arXiv:2605.09595v1 Announce Type: cross Abstract: Reinforcement learning (RL) has enabled robust quadruped locomotion over complex terrain, but most learned controllers are trained offline with backpr

safetyarxiv-cs-ro

NEXUS: Continual Learning of Symbolic Constraints for Safe and Robust Embodied Planning

arXiv:2605.09387v1 Announce Type: new Abstract: While Large Language Models (LLMs) have catalyzed progress in embodied intelligence, a fundamental gap between their inherent probabilistic uncertainty

Not All Turns Matter: Credit Assignment for Multi-Turn Jailbreaking

arXiv:2605.08778v1 Announce Type: new Abstract: Deploying LLMs in multi-turn dialogues facilitates jailbreak attacks that distribute harmful intent across seemingly benign turns. Recent training-based

not every day @scaling01 and I agree. but he’s right. and most people won’t notice it happening, per the work of @informor.

not every day @scaling01 and I agree. but he’s right. and most people won’t notice it happening, per the work of @informor. hot take: unrestricted LLMs are as dangerous as weapons of mass destruction

note that I said “Even @haider1” because he is often optimistic, but he has informed me (with receipts that he shared) that he expressed som…

Gary Marcus references a conversation with someone named Haider, noting that despite Haider's typically optimistic outlook, he has shared documented evidence of expressing concerns about a particular

Now You See That: Learning End-to-End Humanoid Locomotion from Raw Pixels

arXiv:2602.06382v2 Announce Type: replace Abstract: Achieving robust vision-based humanoid locomotion remains challenging due to two fundamental issues: the sim-to-real gap introduces significant perc

safetyarxiv-cs-ro

Offline Preference Optimization for Rectified Flow with Noise-Tracked Pairs

arXiv:2605.09433v1 Announce Type: new Abstract: Existing preference datasets for text-to-image models typically store only the final winner/loser images. This representation is insufficient for rectif

On-Policy Distillation with Best-of-N Teacher Rollout Selection

arXiv:2605.09725v1 Announce Type: new Abstract: On-policy distillation (OPD), which supervises a student on its own sampled trajectories, has emerged as a data-efficient post-training method for impro

On the Generation and Mitigation of Harmful Geometry in Image-to-3D Models

arXiv:2605.09606v1 Announce Type: cross Abstract: Recent advances in image-to-3D models have significantly improved the fidelity and accessibility of 3D content creation. Such a powerful reconstructio

On Uniform Error Bounds for Kernel Regression under Non-Gaussian Noise

arXiv:2605.09757v1 Announce Type: new Abstract: Providing non-conservative uncertainty quantification for function estimates derived from noisy observations remains a fundamental challenge in statisti

On Variance Reduction in Learning Mean Flows

arXiv:2605.09235v1 Announce Type: cross Abstract: One-step generative modeling has emerged as a leading approach to amortize the inference cost of diffusion and flow-matching models. Among distillatio

Open Ontologies: Tool-Augmented Ontology Engineering with Stable Matching Alignment

arXiv:2605.09184v1 Announce Type: new Abstract: We present Open Ontologies, an open-source ontology engineering system implemented in Rust that integrates LLM-driven construction with formal OWL reaso

OpenAI’s Sam Altman’s personal investments are coming under intensifying scrutiny from Republicans following an April article in The Wall St…

Sam Altman's personal investments faced increased scrutiny from Republican lawmakers following reporting by The Wall Street Journal in April regarding potential conflicts of interest. The controversy

OpenClaw-RL: Train Any Agent Simply by Talking

arXiv:2603.10165v2 Announce Type: replace-cross Abstract: Every agent interaction generates a next-state signal, namely the user reply, tool output, terminal or GUI state change that follows each acti

Opinion: It is rare for the US public to agree on anything these days. Fear of AI is as close to a national consensus as it gets. A clear ma…

Opinion: It is rare for the US public to agree on anything these days. Fear of AI is as close to a national consensus as it gets. A clear majority says that AI will do more harm than good. https://ft.

Overcoming Catastrophic Forgetting in Visual Continual Learning with Reinforcement Fine-Tuning

arXiv:2605.09640v1 Announce Type: new Abstract: Recent studies suggest that Reinforcement Fine-Tuning (RFT) is inherently more resilient to catastrophic forgetting than Supervised Fine-Tuning (SFT). H

Path-Coupled Bellman Flows for Distributional Reinforcement Learning

arXiv:2605.08253v1 Announce Type: cross Abstract: Distributional reinforcement learning (DRL) models the full return distribution, but existing finite-support or quantile-based methods rely on project

PATRA: Pattern-Aware Alignment and Balanced Reasoning for Time Series Question Answering

arXiv:2602.23161v2 Announce Type: replace Abstract: Time series reasoning demands both the perception of complex dynamics and logical depth. However, existing LLM-based approaches exhibit two limitati

Pay attention to this one if you build research or knowledge-work agents. Most research-agent systems produce uniform outputs regardless of …

Pay attention to this one if you build research or knowledge-work agents. Most research-agent systems produce uniform outputs regardless of who is driving them. This new work, NanoResearch, argues tha

safetydair-ai--x

Perception Without Engagement: Dissecting the Causal Discovery Deficit in LMMs

arXiv:2605.09422v1 Announce Type: new Abstract: Although Large Multimodal Models (LMMs) have achieved strong performance on general video understanding, their susceptibility to textual prior shortcuts

safetyarxiv-cs-cl

Personalizing LLMs with Binary Feedback: A Preference-Corrected Optimization Framework

arXiv:2605.10043v1 Announce Type: cross Abstract: Large Language Model (LLM) personalization aims to align model behaviors with individual user preferences. Existing methods often focus on isolated us

PersonaTeaming: Supporting Persona-Driven Red-Teaming for Generative AI

arXiv:2605.05682v2 Announce Type: replace-cross Abstract: Recent developments in AI safety research have called for red-teaming methods that effectively surface potential risks posed by generative AI

PFN-TS: Thompson Sampling for Contextual Bandits via Prior-Data Fitted Networks

arXiv:2605.10137v1 Announce Type: cross Abstract: Thompson sampling is a widely used strategy for contextual bandits: at each round, it samples a reward function from a Bayesian posterior and acts gre

PHAGE: Patent Heterogeneous Attention-Guided Graph Encoder for Representation Learning

arXiv:2605.10073v1 Announce Type: new Abstract: Patent claims form a directed dependency structure in which dependent claims inherit and refine the scope of earlier claims; however, existing patent en

safetyarxiv-cs-cl

PHMForge: Evaluating LLM Agents on Industrial Prognostics through MCP-Native, Algorithm-Grounded Tools

arXiv:2604.01532v2 Announce Type: replace Abstract: LLM agents are beginning to invoke industrial asset-management tools through the Model Context Protocol (MCP), yet whether they can act reliably on

PhysEDA: Physics-Aware Learning Framework for Efficient EDA With Manhattan Distance Decay

arXiv:2605.10547v1 Announce Type: new Abstract: Electronic design automation (EDA) addresses placement, routing, timing analysis, and power-integrity verification for integrated circuits. Learning met

Plan in Sandbox, Navigate in Open Worlds: Learning Physics-Grounded Abstracted Experience for Embodied Navigation

arXiv:2605.10118v1 Announce Type: new Abstract: Vision-Language Models (VLMs) have demonstrated exceptional general reasoning capabilities. However, their performance in embodied navigation remains hi

safetyarxiv-cs-ro

Plan2Cleanse: Test-Time Backdoor Defense via Monte-Carlo Planning in Deep Reinforcement Learning

arXiv:2605.09638v1 Announce Type: new Abstract: Ensuring the security of reinforcement learning (RL) models is critical, particularly when they are trained by third parties and deployed in real-world

PMCTS: Particle Monte Carlo Tree Search for Principled Parallelized Inference Time Scaling

arXiv:2605.08982v1 Announce Type: new Abstract: Monte Carlo Tree Search (MCTS) is a widely used approach for policy improvement through search with increasing popularity for real world applications. D

Policy Gradient Methods for Non-Markovian Reinforcement Learning

arXiv:2605.10816v1 Announce Type: cross Abstract: We study policy gradient methods for reinforcement learning in non-Markovian decision processes (NMDPs), where observations and rewards depend on the

Political Plasticity: An Analysis of Ideological Adaptability in Large Language Models

arXiv:2605.08415v1 Announce Type: new Abstract: Since the advent of Large Language Models (LLMs), a significant area of research has focused on their intrinsic biases, particularly in political discou

Position: Academic Conferences are Potentially Facing Denominator Gaming Caused by Fully Automated Scientific Agents

arXiv:2605.09915v1 Announce Type: cross Abstract: The implicit policy of maintaining relatively stable acceptance rates at top AI conferences, despite exponentially growing submissions, introduces a c

Positional Encoding via Token-Aware Phase Attention

arXiv:2509.12635v3 Announce Type: replace-cross Abstract: We prove under practical assumptions that Rotary Positional Embedding (RoPE) introduces an intrinsic distance-dependent bias in attention scor

Positional LSH: Binary Block Matrix Approximation for Attention with Linear Biases

arXiv:2605.09472v1 Announce Type: new Abstract: Positional encoding in transformers is commonly implemented through positional embeddings, attention masks, or bias terms, but formal connections betwee

Positive Alignment: Artificial Intelligence for Human Flourishing

arXiv:2605.10310v1 Announce Type: new Abstract: Existing alignment research is dominated by concerns about safety and preventing harm: safeguards, controllability, and compliance. This paradigm of ali

Primal-Dual Guided Decoding for Constrained Discrete Diffusion

arXiv:2605.09749v1 Announce Type: new Abstract: Discrete diffusion models generate structured sequences by progressively unmasking tokens, but enforcing global property constraints during generation r

Princeton faculty votes to require proctoring in all in-person exams starting this summer, reversing an 1893 policy amid concerns about AI-fueled cheating (Douglas Belkin/Wall Street Journal)

Douglas Belkin / Wall Street Journal: Princeton faculty votes to require proctoring in all in-person exams starting this summer, reversing an 1893 policy amid concerns about AI-fueled cheating — The c

safetytechmeme

Privacy-Aware Video Anomaly Detection through Orthogonal Subspace Projection

arXiv:2605.08651v1 Announce Type: cross Abstract: Video anomaly detection (VAD) systems often prioritize accuracy while overlooking privacy concerns, limiting their suitability for real-world deployme

ProcVLM: Learning Procedure-Grounded Progress Rewards for Robotic Manipulation

arXiv:2605.08774v1 Announce Type: cross Abstract: Long-horizon robotic manipulation requires dense feedback that reflects how a task advances through its procedural stages, not merely whether the fina

PromptGuard: Soft Prompt-Guided Unsafe Content Moderation for Text-to-Image Models

arXiv:2501.03544v5 Announce Type: replace-cross Abstract: Recent text-to-image (T2I) models have exhibited remarkable performance in generating high-quality images from text descriptions. However, the

ProteinOPD: Towards Effective and Efficient Preference Alignment for Protein Design

arXiv:2605.10189v1 Announce Type: cross Abstract: Designing proteins with desired functions or properties represents a core goal in synthetic biology and drug discovery. Recent advances in protein lan

Pseudo-Deliberation in Language Models: When Reasoning Fails to Align Values and Actions

arXiv:2605.09893v1 Announce Type: cross Abstract: Large language models (LLMs) are often evaluated based on their stated values, yet these do not reliably translate into their actions, a discrepancy t

Q-learning with Adjoint Matching

arXiv:2601.14234v3 Announce Type: replace-cross Abstract: We propose Q-learning with Adjoint Matching (QAM), a novel TD-based reinforcement learning (RL) algorithm that tackles a long-standing challen

Quantile-Coupled Flow Matching for Distributional Reinforcement Learning

arXiv:2605.08515v1 Announce Type: new Abstract: Unlike standard expected-return Reinforcement Learning (RL), Distributional RL (DRL) models the full return distribution, making it better-suited for un

Re-Triggering Safeguards within LLMs for Jailbreak Detection

arXiv:2605.10611v1 Announce Type: cross Abstract: This paper proposes a jailbreaking prompt detection method for large language models (LLMs) to defend against jailbreak attacks. Although recent LLMs