Search: “persistence”

Gradient Descent with Large Step Size Restores Symmetry in Deep Linear Networks with Multi-Pathway

arXiv:2606.05219v1 Announce Type: cross Abstract: Recent analyses of multi-pathway Deep Linear Networks use Gradient Flow to predict a 'winner-takes-all' specialization in which path symmetry breaks a

researcharxiv-cs-ai

6 Jun 2026

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Insurance of Agentic AI

DGX agent

arXiv:2606.05449v1 Announce Type: new Abstract: Agentic artificial intelligence (AI) systems are transforming the risk landscape by extending beyond information generation to autonomous planning, tool

agentsarxiv-cs-ai

6 Jun 2026

OPRD: On-Policy Representation Distillation

arXiv:2606.06021v1 Announce Type: cross Abstract: On-policy distillation (OPD) supervises the student only in output space by matching next-token probabilities. This output-only paradigm has two limit

6 Jun 2026

Running Python code in a sandbox with MicroPython and WASM

model-releasessimon-willison

I've been experimenting with different approaches to running code in a sandbox for several years now, but my latest attempt feels like it might finally have all of the characteristics I've been lookin

6 Jun 2026

Alignment Risks from Capability-Seeking RL Training

arXiv:2602.12124v2 Announce Type: replace-cross Abstract: While most AI alignment research focuses on preventing models from generating explicitly harmful content, a more subtle risk arises from capab

safetyarxiv-cs-cl

Benchmarking Open-Source Layout Detection Models for Data Snapshot Extraction from Institutional Documents

arXiv:2606.06242v1 Announce Type: new Abstract: Institutional documents contain substantial amounts of operational and analytical information embedded within figures and tables. Current approaches for

CoT-Space: A Theoretical Framework for Internal Slow-Thinking via Reinforcement Learning

arXiv:2509.04027v3 Announce Type: replace-cross Abstract: Test-time scaling, primarily manifested through multi-step Chain-of-Thought (CoT) reasoning via Reinforcement Learning (RL), has emerged as a

researcharxiv-cs-cl

Do Models Share Safety Representations? Cross-Model Steering for Safe Visual Generation

arXiv:2606.05290v1 Announce Type: new Abstract: Recent progress in generative modeling has made safety control a central challenge, yet existing approaches remain largely model-specific, requiring ret

safetyarxiv-cs-cv

Evaluating Stochastic Collapse and Implicit Bias in Multimodal Large Language Models

arXiv:2606.05874v1 Announce Type: new Abstract: Current evaluations for Multimodal Large Language Models (MLLMs) overwhelmingly focus on utility-driven objectives, leaving model behavior under logic-n

Human Adults and LLMs as Scientists: Who Benefits from Active Exploration?

arXiv:2606.06464v1 Announce Type: new Abstract: A long-standing finding in the causal learning literature is that adults struggle to identify conjunctive causal rules, where an effect requires the sim

safetyarxiv-cs-cl

PlanBench-V: A Spatial Planning Map Benchmark for Vision-Language Models

arXiv:2606.05744v1 Announce Type: new Abstract: Spatial planning maps are central to territorial governance, translating planning objectives, regulations, and spatial strategies into visual forms for

Safe Embodied AI for Long-horizon Tasks: A Cross-layer Analysis of Robotic Manipulation

model-releasesarxiv-cs-ro

arXiv:2606.05660v1 Announce Type: new Abstract: Embodied AI systems are increasingly expected to reason and act over extended horizons in physical environments. This growing capability brings safety t

AlgoVeri: An Aligned Benchmark for Verified Code Generation on Classical Algorithms

arXiv:2602.09464v2 Announce Type: replace-cross Abstract: Vericoding refers to the generation of formally verified code from rigorous specifications. Recent AI models show promise in vericoding, but a

BioBlue: Systematic runaway-optimiser-like LLM failure modes on biologically and economically aligned AI safety benchmarks for LLMs with simplified observation format

arXiv:2509.02655v3 Announce Type: replace-cross Abstract: Many AI alignment discussions of 'runaway optimisation' focus on RL agents: unbounded utility maximisers that over-optimise a proxy objective

Can Generalist Agents Automate Data Curation?

arXiv:2606.04261v1 Announce Type: new Abstract: Curating training data is among the most consequential yet labor-intensive parts of modern AI development: practitioners iteratively propose, implement,

Hardware

Cartridges at Scale: Training Modular KV Caches over Large Document Collections

arXiv:2606.04557v1 Announce Type: new Abstract: Large Language Models can reason over long contexts, yet prefilling millions of tokens is wasteful as much of the content remains static across queries.

hardwarearxiv-cs-cl

Cooperative Circumnavigation for Multiple Unmanned Surface Vehicles Without External Localization

arXiv:2606.04518v1 Announce Type: new Abstract: This paper proposes a cooperative target circumnavigation framework for multiple unmanned surface vehicles (USVs) operating without external localizatio

researcharxiv-cs-ro

Discourse-Role Labels as Presentation-Time Variables for Context Use in Language Models

arXiv:2606.04109v1 Announce Type: new Abstract: Context-augmented language model systems often wrap supplied content with labels such as Reference:, Evidence:, Instruction:, Note:, or Example:, but th

Edge of Stability Selectively Shapes Learning Across the Data Distribution

arXiv:2606.04212v1 Announce Type: new Abstract: Existing analyses of the edge of stability (EoS) treat it as a global property of optimization. We show that it is also selective: the stability constra

safetyarxiv-cs-lg

Flow Matching Calibration for Simulation-Based Inference under Model Misspecification

arXiv:2509.23385v5 Announce Type: replace-cross Abstract: Simulation-based inference (SBI) is transforming experimental sciences by enabling parameter estimation in complex non-linear models from simu

From Prompt to Process: a Process Taxonomy and Comparative Assessment of Frameworks Supporting AI Software Development Agents

arXiv:2606.04967v1 Announce Type: cross Abstract: AI tools for programming are no longer just autocomplete or chat assistants: they organize themselves as development frameworks, with process, roles,

agentsarxiv-cs-ai

LifeSide: Benchmarking Agents as Lifelong Digital Companions

arXiv:2606.04660v1 Announce Type: new Abstract: Lifelong digital companions must integrate cross-session cues, continually update their understanding of users, and adapt to shifting privacy boundaries

Longer Context, Deeper Thinking: Uncovering the Role of Long-Context Ability in Reasoning

arXiv:2505.17315v2 Announce Type: replace Abstract: Recent language models exhibit strong reasoning capabilities, yet the influence of long-context capacity on reasoning remains underexplored. In this

LoopMoE: Unifying Iterative Computation with Mixture-of-Experts for Language Modeling

arXiv:2606.04438v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) and looped architectures scale models along two orthogonal axes, namely parameter capacity and effective depth. However, main

Optimal Transport Flow Matching by Design

arXiv:2606.04092v1 Announce Type: new Abstract: Flow matching models learn to transport samples from a simple prior distribution to a complex data distribution. When prior-data pairs are coupled via o

safetyarxiv-cs-cv

Physics-Informed Video Generation via Mixture-of-Experts Latent Alignment

model-releasesarxiv-cs-cv

arXiv:2606.04737v1 Announce Type: new Abstract: Large-scale video generation models have made remarkable progress in semantic consistency and visual quality, producing videos that are increasingly coh

Preserving Data Privacy in Learning Causal Structure with Fully Homomorphic Encryption

arXiv:2606.05129v1 Announce Type: cross Abstract: Preserving data privacy is an important topic in structural data management and data mining. However, the issue of privacy leakage in distributed caus

researcharxiv-cs-lg

Local Ai

R-APS: Compositional Reasoning and In-Context Meta-Learning for Constrained Design via Reflective Adversarial Pareto Search

arXiv:2606.04823v1 Announce Type: new Abstract: Large language models (LLMs) are fluent on open-ended tasks, yet in agentic settings, where a system must plan, use tools, and act over extended horizon

local-aiarxiv-cs-ai

Revisiting Vul-RAG: Reproducibility and Replicability of RAG-based Vulnerability Detection with Open-Weight Models

arXiv:2606.04739v1 Announce Type: cross Abstract: Large language models (LLMs) have shown strong potential for automated software vulnerability detection, particularly in retrieval-augmented generatio

Sometimes you need to start over. But that decision is hard. @llama_index had to make that call: they built one of the most popular AI frame…

Sometimes you need to start over. But that decision is hard. @llama_index had to make that call: they built one of the most popular AI frameworks in the world, but saw the agent harness and frontier l

agentsjerry-liu--x

Toward a Generalized Defense Across Sparse, Continuous, and Structured Parameter Attacks

arXiv:2606.04317v1 Announce Type: cross Abstract: Deep neural networks are increasingly deployed across heterogeneous and partially untrusted environments, where models are distributed through cloud s

Transferable Multi-Bit Watermarking Across Frozen Diffusion Models via Latent Consistency Bridges

arXiv:2603.20304v2 Announce Type: replace Abstract: As generative AI advances, global governance frameworks increasingly mandate verifiable content provenance. However, existing watermarking technique

safetyarxiv-cs-cv

Applications

UltraEP: Unleash MoE Training and Inference on Rack-Scale Nodes with Near-Optimal Load Balancing

arXiv:2606.04101v1 Announce Type: cross Abstract: Large-scale expert parallelism (EP) is becoming pivotal for training and serving frontier MoE models, but it also amplifies device-level expert load i

applicationsarxiv-cs-lg

A 3D Isovist World Model -- Revealing a City's Unseen Geometry and Its Emergent Cross-City Signature

arXiv:2606.03609v1 Announce Type: cross Abstract: Embodied agents that navigate cities rely on world models that predict how their surroundings will change as they move. But for navigation, what matte

AI Model Extraction Attacks: Bypassing Single-Client Assumptions in Defenses

arXiv:2606.03381v1 Announce Type: cross Abstract: Ensuring the protection of Artificial Intelligence (AI) models deployed in military Command and Control (C2) systems and critical infrastructure is es

researcharxiv-cs-ai

Are Deep Learning Based Hybrid PDE Solvers Reliable? Why Training Paradigms and Update Strategies Matter

arXiv:2602.06842v2 Announce Type: replace-cross Abstract: Deep learning-based hybrid iterative methods (DL-HIMs) integrate classical numerical solvers with neural operators, utilizing their complement

researcharxiv-cs-lg

DeskCraft: Benchmarking Desktop Agents on Professional Workflows and Human-in-the-Loop Collaboration

arXiv:2606.03103v1 Announce Type: new Abstract: Real-world professional desktop workflows in specialized creative and engineering software unfold over long horizons and often require human-in-the-loop

Do Neural Retrievers Prefer Certain Documents? Evidence of Learned Relevance Priors

arXiv:2606.02814v1 Announce Type: cross Abstract: Neural retrievers are trained to estimate query-document relevance from annotated query-document pairs. Yet annotation protocols may not purely reflec

safetyarxiv-cs-ai

Do Value Vectors in Deep Layers Need Context from the Residual Stream?

arXiv:2606.02780v1 Announce Type: new Abstract: The success of the transformer architecture as the backbone of modern LLMs is in large part due to its use of attention layers. An attention layer follo

Don't Forget Your Embeddings: Robust Knowledge Erasure via Precise Editing of Embeddings

arXiv:2606.03695v1 Announce Type: new Abstract: As language models are increasingly deployed in real-world applications, the ability to erase specific knowledge from them becomes critical for safety a

FinStressTS: A Parametric Synthetic Benchmark for Time-Series Forecasting in Finance

arXiv:2606.03184v1 Announce Type: cross Abstract: Financial forecasting is difficult due to low signal-to-noise ratios, latent factors, heavy tails, regime shifts, and jumps. Real-world benchmarks off

Local Ai

Framing Migration News with LLMs: Structured CoT as a Support for Human Interpretation

arXiv:2606.03761v1 Announce Type: new Abstract: Frame analysis of migration news is a socially consequential task: media scholars and researchers who study how migration is narrated need tools that ar

local-aiarxiv-cs-cl

From Answers to States: Verifiable Process-Level Evaluation of Chemical Reasoning in Large Language Models

arXiv:2606.03660v1 Announce Type: new Abstract: Large language models are increasingly used as chemistry assistants, yet most chemistry benchmarks still score only final answers. This masks a critical

Gender-Dependent Diagnostic Substitution in LLM Medical Triage: Same Symptoms, Unequal Urgency

arXiv:2606.03641v1 Announce Type: new Abstract: We investigate whether large language models produce different medical triage recommendations for identical neurological symptoms when only the patient'

Hallucinations as Orthogonal Noise: Inference-Time Manifold Alignment via Dynamic Contextual Orthogonalization

arXiv:2606.03022v1 Announce Type: cross Abstract: Hallucination in Large Language Models (LLMs), characterized by the generation of content inconsistent with contextual facts or logical constraints --

KForge: LLM-Driven Cross-Platform Kernel Generation for AI Accelerators

arXiv:2606.02963v1 Announce Type: new Abstract: Production inference increasingly targets a heterogeneous mix of accelerators. Agentic pipelines interleave reasoning, tool calls, and multi-agent coord

Large AI Models in Dental Healthcare: From General-Purpose Systems to Domain-Specific Foundation Models

arXiv:2606.02914v1 Announce Type: new Abstract: Background: Oral diseases affect nearly 3.5 billion people worldwide, yet the comparative clinical potential of large-scale AI models in dentistry remai

agentsarxiv-cs-ai

Let There Be Light: Reflection, Refraction and Scattering for Neural Operators

arXiv:2606.03262v1 Announce Type: new Abstract: Neural operators learn mappings between infinite-dimensional function spaces and provide a data-driven surrogate modeling paradigm for parametric partia

safetyarxiv-cs-lg