Search: “safety”

The Easy, the Hard, and the Learnable: Confidence and Difficulty-Adaptive Policy Optimization for LLM Reasoning

arXiv:2606.07950v1 Announce Type: new Abstract: RL with verifiable rewards can substantially improve LLM reasoning, yet standard GRPO-style training often treats easy, hard, and learnable questions al

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

The Flexibility Trap: Rethinking the Value of Arbitrary Order in Diffusion Language Models

DGX agent

arXiv:2601.15165v4 Announce Type: replace-cross Abstract: Diffusion Large Language Models (dLLMs) break the rigid left-to-right constraint of traditional LLMs, enabling token generation in arbitrary o

safetyarxiv-cs-ai

9 Jun 2026

The Hidden Bias of Process Reward Models:PRISM for Rewarding the Right Reasoning

arXiv:2606.09078v1 Announce Type: new Abstract: Process Reward Models (PRMs) improve credit assignment for reasoning by providing step-level feedback. However, we identify a hidden bias in PRMs caused

The Spectral Dynamics and Noise Geometry of Muon

arXiv:2606.08388v1 Announce Type: new Abstract: Muon replaces a matrix gradient G=USigma V^op by its polar factor UV^op. This keeps the singular directions selected by the gradient, but makes the upda

Think Before You Act: Intention-Guided Reasoning for LLM-Based Location Prediction

arXiv:2606.08122v1 Announce Type: new Abstract: Predicting a user's next Point-of-Interest (POI) based on their historical check-in records is a fundamental task in location-based services. While rece

This is actually a pretty measured bet compared to spending of the U.S companies. OpenAI alone is about twice as much. If things falls apart…

This is actually a pretty measured bet compared to spending of the U.S companies. OpenAI alone is about twice as much. If things falls apart, the US will be hit harder. 🚨BREAKING: CHINA IS AGI-PILLED

This quote from OpenAI is telling. Translation: We have a rapidly closing window to get this IPO out the door before the bubble bursts, our …

This quote from OpenAI is telling. Translation: We have a rapidly closing window to get this IPO out the door before the bubble bursts, our CFO doesn't want to go through with it because our books are

TinyJudge: Unverifiable Constraint Alignment via Lightweight Specialist Ensembles

arXiv:2606.07520v1 Announce Type: cross Abstract: Instruction Following (IF) is a core capability of LLMs, requiring strict adherence to diverse constraints, ranging from verifiable ones (e.g., output

TORL-VLA: Tactile Guided Online Reinforcement Learning for Contact-Rich Manipulation

arXiv:2606.09337v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models have become a powerful framework for robotic manipulation, and recent studies have introduced tactile or force feedb

Towards Accurate Emotion-Attributed Video Captioning via Fine-grained Emotion-Cause Pair Extraction

arXiv:2606.08566v1 Announce Type: new Abstract: Emotional Video Captioning (EVC) is a challenging task that aims to generate factually accurate and emotionally rich descriptions for videos. Existing E

Towards End to End Motion Planning and Execution for Autonomous Underwater Vehicles Using Reinforcement Learning

arXiv:2606.08513v1 Announce Type: cross Abstract: Autonomous Underwater Vehicles (AUVs) traditionally rely on complex, heavily engineered pipelines for perception, path planning, and motion control. T

Towards Mitigating Hallucinations in Large Vision-Language Models by Refining Textual Embeddings

arXiv:2511.05017v2 Announce Type: replace Abstract: Hallucinations in Large Vision-Language Models (LVLMs) remain a persistent challenge, often stemming from inadequate integration of visual informati

Training-Inference Kernel Contracts: Bounding Divergence in Post-Training and Deployment

arXiv:2606.07581v1 Announce Type: cross Abstract: A modern post-training pipeline often writes one symbol for its policy, pi_theta, while evaluating it through two different programs: a training kerne

Trait-space Monitoring for Emergent Misalignment During Supervised Finetuning

arXiv:2606.07631v1 Announce Type: cross Abstract: Emergent misalignment (EM) occurs when narrow finetuning causes a model to behave dangerously outside the finetuning task. Standard training signals c

TRUST-SCF: Transformer-based Risk Understanding and Scoring for Transactional Supply Chain Finance

arXiv:2606.08140v1 Announce Type: new Abstract: Supply Chain Finance (SCF) and LendTech platforms need credit scoring systems that respond to evolving transaction behavior, repayment delays, and activ

Two Bridges, One Pathway: From VLMs to Generalizable VLAs with Embodied Trajectory-Coupled Data

arXiv:2606.08520v1 Announce Type: new Abstract: Vision-language models (VLMs) are powerful general-purpose reasoners, yet converting them into robot control policies (VLAs) is surprisingly difficult.

Uncertainty-Aware Hierarchical Re-Localization in OpenStreetMap via Semantic Alignment

arXiv:2603.01613v2 Announce Type: replace Abstract: Monocular re-localization enables robots to estimate camera poses from visual observations. However, many existing methods rely on dense maps or lar

Unifying Object-Centric World Models and Diffusion Policy: A Hierarchical Framework for Multi-Stage Robotic Tasks

arXiv:2606.08775v1 Announce Type: cross Abstract: Visual world models have shown great potential in learning complex system dynamics. Recent advancements leverage these models as transition functions

VAIC: Vision-Guided Humanoid Agile Object Interaction Control via Decoupled Commands

arXiv:2606.09286v1 Announce Type: new Abstract: Humanoid robots hold immense potential for real-world assistance, yet agile interaction with objects in unstructured environments demands tightly couple

Video Understanding by Design: How Datasets Shape Video Models

arXiv:2509.09151v2 Announce Type: replace-cross Abstract: Research in video understanding has advanced rapidly, driven by increasingly diverse datasets and more powerful model architectures. While exi

Vision-Language Asymmetry in Bistable Image Captioning

arXiv:2606.08031v1 Announce Type: new Abstract: Wittgenstein's duck-rabbit poses a question for vision-language models: when a model captions an ambiguous image, where in the model is the commitment t

Visual Para-Thinker++: A Single-Policy Multi-Agent Framework for Visual Reasoning

arXiv:2606.09290v1 Announce Type: new Abstract: Visual reasoning requires integrating evidence distributed across regions, attributes, and relations, making single-chain reasoning prone to early perce

wait til you see what happens with SpaceX.

Gary Marcus, a prominent AI researcher and critic, posted a teaser statement on X about upcoming SpaceX developments, likely referring to significant announcements or milestones in the company's space

WaveDiT: Distribution-Aware Wavelet Flow Matching for Efficient 3D Brain MRI Synthesis

arXiv:2606.08670v1 Announce Type: new Abstract: Large and demographically balanced datasets are essential for reliable neuroimaging biomarkers. Full-resolution 3D brain MRI synthesis can support data

Web Agents Should Use Typed Actions Instead of Click-Based Browsing

arXiv:2602.17245v2 Announce Type: replace Abstract: This position paper argues that building a reliable agentic Web requires shifting from low-level interaction primitives to typed actions supported b

what constitutes reasoning in AI is a critical debate. i hope that @dwarkesh_sp will respond.

what constitutes reasoning in AI is a critical debate. i hope that @dwarkesh_sp will respond. what does this even mean, @dwarkesh_sp, “the real deal”? is it even a falsifiable conjecture? what’s the e

what does this even mean, @dwarkesh_sp, “the real deal”? is it even a falsifiable conjecture? what’s the evidence? and if you can agree that…

what does this even mean, @dwarkesh_sp, “the real deal”? is it even a falsifiable conjecture? what’s the evidence? and if you can agree that “If you ask an LLM a question it can't answer, sometimes it

What Makes a Desired Graph for Relational Deep Learning?

arXiv:2606.08491v1 Announce Type: new Abstract: Relational deep learning (RDL) converts relational databases (RDBs) into heterogeneous graphs, but graphs derived directly from database schemas are oft

WhiFlash: Accelerating Speculative Decoding with Token-Level Cross-Paradigm Routing

arXiv:2606.07710v1 Announce Type: cross Abstract: The autoregressive nature of large language models (LLMs) remains a significant bottleneck for inference, particularly in complex agentic workloads. W

Why are governments in the dark about the threat posed by superintelligent AI? Because the industry is spending to keep them there. In ABC's…

Why are governments in the dark about the threat posed by superintelligent AI? Because the industry is spending to keep them there. In ABC's new Four Corners documentary, ControlAI's US Director Conno

safetyconnor-leahy--x

X-OP: Cross-Morphology Whole-Body Teleoperation via MPC Retargeting

arXiv:2606.07934v1 Announce Type: new Abstract: Whole-body teleoperation is essential for scalable robot data collection in loco-manipulation tasks, yet existing approaches relying on exoskeleton suit

you gotta love these juxtapositions.

Gary Marcus highlights interesting contrasts or contradictions, likely related to AI, technology, or cognitive science based on his expertise in these areas. The post format suggests he's pointing out

Your Self-Play Algorithm is Secretly an Adversarial Imitator: Understanding LLM Self-Play through the Lens of Imitation Learning

arXiv:2602.01357v2 Announce Type: replace Abstract: Self-play post-training methods has emerged as an effective approach for finetuning large language models and turn the weak language model into stro

A federal judge blocks President Trump's policy imposing a $100,000 fee for employer H-1B visa applications (CNBC)

CNBC: A federal judge blocks President Trump's policy imposing a $100,000 fee for employer H-1B visa applications — Choose CNBC as your preferred source on Google and never miss a moment from the most

safetytechmeme

A Geometric Account of Activation Steering through Angle-Norm Decomposition

arXiv:2606.06735v1 Announce Type: new Abstract: Linear activation steering has gained popularity as a simple and empirically effective way to control language model behavior. More recently, spherical

A Geometric View for Understanding Concept Learning and Neuron Interpretation in Sparse Autoencoders

arXiv:2606.07007v1 Announce Type: cross Abstract: We propose a unified mathematical framework for a geometric understanding of concept learning and neuron interpretation in sparse autoencoders (SAEs).

A look at Ireland's Bring Your Own Power policy, which requires new data centers to have their own power plants on site or sources of energy produced nearby (Yusuf Khan/Wall Street Journal)

Yusuf Khan / Wall Street Journal: A look at Ireland's Bring Your Own Power policy, which requires new data centers to have their own power plants on site or sources of energy produced nearby — The tin

safetytechmeme

a lot of fools and a lot of money shall soon be parted.

Gary Marcus, a prominent AI researcher and critic, posted a commentary suggesting that foolish investors and those with excessive capital will soon experience significant financial losses, likely refe

A pause continues to be utter and complete nonsense and it always will be. 1. Let's make planes safer by not making planes! 2. What exactly …

A pause continues to be utter and complete nonsense and it always will be. 1. Let's make planes safer by not making planes! 2. What exactly happens in a 'pause'? Do labs get to keep working and we all

safetyyann-lecun--x

Accounting for Context: Shaping Moral Credences for Value Alignment

arXiv:2606.06972v1 Announce Type: new Abstract: Ensuring that agent behaviours are aligned with human moral values inevitably raises the problem of how to account for the plurality of moral perspectiv

Acoustic Cue Alignment in Audio Language Models for Speech Emotion Recognition

arXiv:2606.07309v1 Announce Type: cross Abstract: Instruction-following audio language models (ALMs) can be augmented with explicit acoustic cues, yet it remains unclear whether such cues are used in

ActionMap: Robot Policy Learning via Voxel Action Heatmap

arXiv:2606.06904v1 Announce Type: cross Abstract: Vision-language-action (VLA) models have advanced rapidly across backbones, training recipes, and data scale, yet the action decoder, which converts t

AdaGRPO: A Capability-Aware Adaptive Enhancement for Flow-based GRPO

arXiv:2606.06828v1 Announce Type: new Abstract: Group Relative Policy Optimization (GRPO) has demonstrated remarkable success in aligning text-to-image (T2I) flow models with human preferences. Howeve

AdaJudge: Adaptive Multi-Perspective Judging for Reward Modeling

arXiv:2601.08097v2 Announce Type: replace Abstract: Reward modeling is essential for aligning large language models with human preferences, yet predominant architectures rely on a static pooling strat

safetyarxiv-cs-cl

AdaTok: Self-Budgeting Image Tokenization with Quality-Preserving Dynamic Tokens

arXiv:2606.07185v1 Announce Type: new Abstract: Image tokenizers, from 2D grids to recent 1D sequences, typically encode every image with the same fixed number of tokens. Yet visual complexity is high

AEGIS: A Backup Reflex for Physical AI

arXiv:2606.06660v1 Announce Type: new Abstract: Long-horizon robot manipulation tends to fail gradually: one bad step degrades the state, and the policy spirals into a basin from which it cannot recov

Affordance-Based Hierarchical Reinforcement Learning for Quadruped Pedipulation

arXiv:2606.07506v1 Announce Type: new Abstract: The object manipulation capabilities of quadruped robots is an open research challenge. While previous studies have focused on low-level policy learning

Agentic World Modeling for 6G: Near-Real-Time Generative State-Space Reasoning

arXiv:2511.02748v2 Announce Type: replace-cross Abstract: We argue that sixth-generation (6G) intelligence is not fluent token prediction but the capacity to imagine and choose -- to simulate future s