Search: “models”

Attention Routing Stabilizes Early: Working-Set Inference for Recurrent Language Models

arXiv:2609.27373v1 Announce Type: new Abstract: Recurrent language models repeatedly apply shared network blocks to refine latent representations, but standard inference recomputes global attention at

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

BEE: Intervention-Adaptive Real-World Reinforcement Learning with Vision-Language-Action Models

DGX agent

arXiv:2609.27450v1 Announce Type: cross Abstract: Vision-language-action (VLA) models handle long-horizon manipulation, yet success hinges on a few precision-critical phases where millimeter-scale err

safetyarxiv-cs-ai

24 Sept 2026

Hardware

Efficient MoE Training for Biological Foundation Models

Mixture‑of‑Experts (MoE) models scale capacity more efficiently than dense transformers by activating only a subset of expert subnetworks per token. NVIDIA’s Transformer Engine optimizes MoE training

hardwarenvidia-developer

Groundbench: Multi-Resolution Polygon Grounding Exposes the Geometry Gap in Vision-Language Models

model-releasesarxiv-cs-cv

arXiv:2609.27821v1 Announce Type: new Abstract: Bounding-box scores on RefCOCO-family grounding leave little room to distinguish frontier vision-language systems, yet boxes discard object shape. We in

Agents

It's not 'superintelligence' that's dangerous, but the business models of OpenAI & Anthropic. They want businesses to use autonomous agents …

It's not 'superintelligence' that's dangerous, but the business models of OpenAI & Anthropic. They want businesses to use autonomous agents because agents chew up more tokens. But 'agency' =/= 'intell

agentsgary-marcus--x

LOCKR: A Hidden-State Trajectory-Guided Planner for Detecting and Repairing Stable-but-Wrong Lock-In in Diffusion Language Models

arXiv:2609.27220v1 Announce Type: new Abstract: Diffusion language models generate text through iterative denoising, exposing intermediate trajectories before final answers are produced. We identify a

MemBodied: Recurrent Associative Memory for Vision-Language-Action Models

arXiv:2609.28256v1 Announce Type: cross Abstract: Vision-Language-Action models provide a strong foundation for general-purpose robot control, yet a vast majority of policies do not preserve and lever

safetyarxiv-cs-ai

Robustness of Diffusion Models under Distribution Shift

arXiv:2609.27546v1 Announce Type: cross Abstract: Score-based diffusion models are increasingly considered in settings where the underlying data distribution may differ from the training distribution,

researcharxiv-cs-lg

Industry

Sources: the White House asked OpenAI and Anthropic not to share new models with UK's AISI until US reviews them; Anthropic appears to have agreed (Politico)

Politico: Sources: the White House asked OpenAI and Anthropic not to share new models with UK's AISI until US reviews them; Anthropic appears to have agreed — The White House has asked OpenAI and Anth

industrytechmeme

The Gaussian Is Enough: Flow-Matching Priors Do Not Help When Fine-Tuning Large Behavior Models

arXiv:2609.27070v1 Announce Type: cross Abstract: Modern robot imitation learning increasingly relies on generative policies based on diffusion or flow-matching models, which generate actions by trans

Tools

Today, @FactoryAI joins the Specialized Intelligence Index with Legacy-Bench! The SII keeps growing, now with benchmarks testing how models …

Today, @FactoryAI joins the Specialized Intelligence Index with Legacy-Bench! The SII keeps growing, now with benchmarks testing how models handle the legacy software behind payroll, financial settlem

toolsfireworks-ai--x

Beyond Reconstruction Error: Analytical and Data-Driven Action Tokenization for Autoregressive Vision-Language-Action Models

arXiv:2609.25820v1 Announce Type: cross Abstract: Discrete action tokenization is central to autoregressive vision-language-action (VLA) models, yet action representations are often evaluated primaril

safetyarxiv-cs-lg

Deploying Foundation Models for Embodied Navigation

arXiv:2609.25666v1 Announce Type: new Abstract: We present and tackle two problems associated with deploying Foundation Models (FMs) on Embodied Agents performing navigation: 1) Training bias in FMs l

safetyarxiv-cs-ro

Dual-Locking Learned AI Models: A PIN-Based Sparse QIM Watermarking and Adaptive Index Permutation Approach

arXiv:2609.22981v1 Announce Type: cross Abstract: We present a dual-locking method for securing trained neural networks that combines key-driven index permutation with PIN-based watermarking based on

safetyarxiv-cs-ai

PACE-dLLM: Elastic Block Decoding via Confidence Cliff Estimation for Diffusion Language Models

arXiv:2609.26249v1 Announce Type: new Abstract: Diffusion language models (dLLMs), such as LLaDA and Dream, have become competitive with autoregressive (AR) LLMs in generation quality while supporting

Local Ai

Pretraining of Medical Visual Encoders Toward Multi-modal Large Language Models

arXiv:2609.23860v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) commonly reuse visual encoders pretrained with CLIP, although the features of these ViTs are ultimately consume

local-aiarxiv-cs-ai

Same Quantity, Different Answer: Numerical Representation Invariance in Language Models

model-releasesarxiv-cs-cl

arXiv:2609.25009v1 Announce Type: new Abstract: Numerically equivalent word problems should yield the same canonical answer whether a quantity is written as a decimal, fraction, percentage, number wor

Skytopia: Monocular Drone Navigation with Action-Conditioned Latent World Models

arXiv:2609.26007v1 Announce Type: new Abstract: Monocular drone navigation requires reaching a goal in an unseen environment from a single forward-facing camera, which offers few cues for depth and sc

safetyarxiv-cs-ro

Spectral Tail Interventions in Decoder-Only Language Models: Reasoning-Sensitive Weight Structure from Controlled Surgery

arXiv:2609.26165v1 Announce Type: new Abstract: Weight-space structure often correlates with language-model behavior, but correlation alone does not establish computational involvement. We study conce

researcharxiv-cs-lg

The Challenge of Identifying the Origin of Black-Box Large Language Models

arXiv:2503.04332v2 Announce Type: replace-cross Abstract: The tremendous commercial potential of large language models (LLMs) has heightened concerns over their unauthorized use. To address this, we f

applicationsarxiv-cs-lg

The Probabilistic Structure of Large Language Models

arXiv:2609.25134v1 Announce Type: new Abstract: This paper presents a probabilistic perspective on large language models (LLMs), developed with the aim of bringing together, in a single self-contained

researcharxiv-cs-lg

Agents

TypeSafe is pushing hard on this idea of “composable AI” but LangChain has embodied composable AI since day one. With new models like Jev, w…

TypeSafe is pushing hard on this idea of “composable AI” but LangChain has embodied composable AI since day one. With new models like Jev, we’re decomposing the agent loop and moving more intelligence

agentsharrison-chase--x

Video-HopChain: Multi-Hop Questions and Confidence-Gated Exploration for Video Reasoning Models

model-releasesarxiv-cs-cv

arXiv:2609.25773v1 Announce Type: new Abstract: HopChain has shown on still images that multi-hop data synthesis improves vision-language reasoning, because long chain-of-thought reasoning exposes err

Agents

A Survey of Agentic Reasoning for Large Language Models: Towards Recursively Self-Improving and Collective Agents

arXiv:2601.12538v2 Announce Type: replace-cross Abstract: Reasoning is a fundamental cognitive process underlying inference, problem-solving, and decision-making. While large language models (LLMs) de

agentsarxiv-cs-cl

Contextual Causality with Large Language Models: A Survey

arXiv:2609.22409v1 Announce Type: new Abstract: Understanding contextual causality is critical for large language models (LLMs), as it enables them to accurately identify causal relations in specific

applicationsarxiv-cs-cl

Didactic knowledge or Clinical Cases? How Data Types Shape Medical Large Language Models

arXiv:2609.22161v1 Announce Type: cross Abstract: Medical large language models are commonly trained on mixtures of didactic data (e.g., textbooks) and clinical data (e.g., patient records), yet how t

Diff-2-in-1: Bridging Generation and Dense Perception with Diffusion Models

arXiv:2411.05005v2 Announce Type: replace Abstract: Beyond high-fidelity image synthesis, diffusion models have recently exhibited promising results in dense visual perception tasks. However, most exi

researcharxiv-cs-cv

From Trait Vectors to Circuits: Tracing Refusal and Sycophancy Through Language Models

arXiv:2609.22224v1 Announce Type: new Abstract: A direction in activation space that changes safety-relevant behavior when steered is not necessarily one the model uses to produce that behavior on its

Functional Emotion Without Character: Large Language Models, Aristotelian Disposition, and the Limits of Behavioral Alignment

arXiv:2609.22362v1 Announce Type: new Abstract: Debates about whether artificial systems can feel are often forced between two unsatisfactory positions: behavioral equivalence is treated as sufficient

Industry

GPT-6 Sol and Luna are great models but also these characters are so cute

GPT-6 Sol and Luna are great models but also these characters are so cute Please welcome GPT-6 Sol and GPT-6 Luna to the GPT-6 universe. GPT-6 Sol and Luna build on the advances behind GPT-6 Astra, br

industrysam-altman--x

INTCORT: Training-Free Spatial Reasoning Enhancement for Vision-Language Models via Input Transformations and Confidence Routing

arXiv:2609.24813v1 Announce Type: new Abstract: Vision-Language Models (VLMs) have demonstrated remarkable capabilities in multimodal tasks, yet they still exhibit poor ability in spatial reasoning. E

researcharxiv-cs-cv

Large language models in medical time series analysis

arXiv:2609.22262v1 Announce Type: cross Abstract: Medical time series (MedTS), including electrocardiograms (ECG), electroencephalograms (EEG), photoplethysmography (PPG), and vital-sign recordings, a

applicationsarxiv-cs-lg

Monotone-Constrained Diffusion Models for Long-Horizon Production Forecasting

model-releasesarxiv-cs-lg

arXiv:2609.22643v1 Announce Type: new Abstract: Forecasting a long horizon from only the first observations of a sequence is ill-posed: many trajectories are consistent with the same short history. We

OpenAI says it plans to let third-party groups conduct technical safety evaluations of its AI models during the training, evaluation, and deployment phases (Rachel Metz/Bloomberg)

Rachel Metz / Bloomberg: OpenAI says it plans to let third-party groups conduct technical safety evaluations of its AI models during the training, evaluation, and deployment phases — OpenAI plans to l

safetytechmeme

Local Ai

Pharmacokinetic State Space Models for Unbiased Prediction of Haemodynamic Collapse

arXiv:2609.24338v1 Announce Type: new Abstract: An Intraoperative Hypotension (IOH) event is a frequent complication during administration of general anaesthesia with serious downstream consequences,

local-aiarxiv-cs-lg

Rethinking Pivot Programming Languages in Code Language Models

arXiv:2609.22988v1 Announce Type: new Abstract: Multilingual code language models transfer skills across programming languages (PLs), but whether any PL occupies a privileged pivot position remains co

Token Signatures of Code: Comparing Coding Behaviors Across Large Language Models

arXiv:2609.22097v1 Announce Type: new Abstract: The evaluation of large language models (LLMs) on coding tasks has primarily focused on performance metrics such as pass@k. As LLMs continue to advance,

When Cosine Similarity Fails to Reflect Linearly Accessible Structure in Dialogue Models

arXiv:2609.22522v1 Announce Type: new Abstract: Cosine similarity is widely used to analyze transformer representations, implicitly assuming that similarity reflects task-relevant structure. We study

A Generative Grammar Underlying the Voynich Manuscript, the Pastiche Hypothesis: Evidence from Large Language Models

arXiv:2609.20835v1 Announce Type: new Abstract: Background: The Voynich Manuscript is a fifteenth-century codex written in an unknown script whose content remains undeciphered. Previous studies sugges

DiaVLo: Diagnosing Behaviours of Vision-Language Models

arXiv:2609.22008v1 Announce Type: cross Abstract: Vision-language models (VLMs) rely on storing and transferring appropriate information across their sub-components. Verifying that the VLMs exhibit de

Learning to Move Cities: Deep Meta-Models and Reinforcement Policies for Calibration and Control in Urban Networks

model-releasesarxiv-cs-lg

arXiv:2609.21945v1 Announce Type: new Abstract: Urban transportation networks present complex optimization challenges spanning calibration of high-fidelity simulators and real-time operational control

LogicTrack: Auditing Reasoning Trajectories of Large Language Models with Formal Logic Solvers

arXiv:2609.21492v1 Announce Type: new Abstract: Chain-of-Thought (CoT) reasoning has been shown to improve the performance of large language models (LLMs), yet existing optimization methods largely re

Hardware

RBS-Attention: Radius-Bounded Sparse Prefill for Long-Context Large Language Models

arXiv:2609.20971v1 Announce Type: new Abstract: Long-context large language model inference is increasingly limited by prefill, where dense self-attention processes the entire prompt before generation

hardwarearxiv-cs-ai

Recursive Language Models Generalize Out of Domain

arXiv:2609.20831v1 Announce Type: new Abstract: We study when limiting what a language model can see improves learning. We compare standard CoT, the more general learner that reads the full trace, wit

Rewarding Efficient Reasoning Improves Abstention on Underspecified Tasks in Reasoning Models

arXiv:2609.20846v1 Announce Type: new Abstract: While modern large reasoning models (LRMs) excel at providing correct answers in many tasks, we provide additional evidence for the observation that the

CleanVideo: Adaptive Concept Erasure for Text-to-Video Diffusion Models

arXiv:2609.20267v1 Announce Type: cross Abstract: Concept erasure aims to selectively eliminate undesired visual semantics from pre-trained generative models without compromising their general utility

18 Sept 2026

Compositional Reasoning in Language Models under Reinforcement Learning Post-Training

arXiv:2609.19465v1 Announce Type: new Abstract: Compositional reasoning is critical for real-world problem solving: since training data is necessarily limited, models must generalize by composing lear

applicationsarxiv-cs-ai

18 Sept 2026

MoWAM: Explicit Future Motion Prediction for Efficient World Action Models