Search: “llm”

Sparse Attention Is Matrix Approximation, Not Choosing from a Bag of Values

arXiv:2610.10871v1 Announce Type: new Abstract: Large Language Models (LLMs) achieve strong performance across many domains, but their efficiency is limited by the quadratic cost of attention with res

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Spectral Weight Decay: Inducing Low-Rank Structure in Neural Network Weights

DGX agent

arXiv:2610.11730v1 Announce Type: new Abstract: Standard weight decay treats each weight matrix as a vector and ignores its spectral structure. We introduce spectral weight decay, a post-step decouple

model-releasesarxiv-cs-lg

9 Oct 2026

Stop Probing, Start Coding: Why Linear Probes and Sparse Autoencoders Fail at Compositional Generalisation

model-releasesarxiv-cs-lg

arXiv:2603.28744v2 Announce Type: replace Abstract: The linear representation hypothesis states that neural network activations encode high-level concepts as linear mixtures. However, under superposit

Synthesis Through Simulation: Generating Coherent Enterprise Data via Scalable Agent-System Interaction

arXiv:2610.10549v1 Announce Type: new Abstract: Tool-calling agents have become central to enterprise AI, yet training and evaluating them at scale remains severely constrained due to business and leg

safetyarxiv-cs-ai

The '10th Juror': Open-Set Standpoint Screening for Bureaucratic Bias Detection

arXiv:2610.11136v1 Announce Type: new Abstract: Presupposing the boundaries of bias is itself a form of bias. We study closed-loop bias governance for Dutch government documents, where a system must d

local-aiarxiv-cs-cl

ThinkingBox: Solving an agent task once vs. solving it 20/20: 507 stateful workflows graded on terminal database state [R]

model-releasesr-machinelearning

Figure 1b from our paper Disclosure: I'm one of the authors (Microsoft). The paper, code, dataset are public and ThinkingBox is on Hugging Face OpenEnv as well. Raw evaluation trajectories are not rel

TRL v1.15 is out, and it’s an absolute banger of a release for memory-efficient post-training. The main change: SFT, DPO, KTO, GRPO, RLOO an…

model-releasesclem-delangue--x

TRL v1.15 is out, and it’s an absolute banger of a release for memory-efficient post-training. The main change: SFT, DPO, KTO, GRPO, RLOO and Distillation now use a fused LM head by default. Instead o

Tutorials

USDCraft: Geometrically Grounded Programmatic Modeling of Articulated 3D Assets for Simulation

arXiv:2610.11322v1 Announce Type: new Abstract: Geometrically faithful and functional articulated 3D assets are essential for real-to-sim robot manipulation, where policies trained in simulation must

tutorialsarxiv-cs-ro

When Interfaces Speak: Data-Aware Generative UI Harness for Active Interaction

model-releasesarxiv-cs-ai

arXiv:2610.11123v1 Announce Type: new Abstract: Most human-agent interaction today remains text-based. Natural language can impose cognitive overload, ambiguity, information chaos, and slow input for

When Scene Text Hijacks the Scene: Uncovering, Exploiting, and Mitigating Rendered-Text Semantic Leakage in Image Generation Models

arXiv:2610.11286v1 Announce Type: new Abstract: The reliability and accountability of image generative models (IGMs) are essential for building responsible and trustworthy AI systems. Recent IGMs, suc

local-aiarxiv-cs-cv

Adaptive Code Generation for Controlling Robots

arXiv:2610.09588v1 Announce Type: new Abstract: Deploying robots as Complex Adaptive Systems (CAS) in unknown and dynamic environments necessitates a transition from rigid command libraries toward int

researcharxiv-cs-ro

AeroEval: Staged Program and Execution Validation for AI-Generated Drone Missions

arXiv:2610.09764v1 Announce Type: new Abstract: Large Language Models (LLMs) can generate drone programs from natural-language mission descriptions, but syntactically valid programs may still violate

local-aiarxiv-cs-ro

An Empirical Study of Agent Skills' Downstream Utility

arXiv:2610.08875v1 Announce Type: cross Abstract: Agent Skills package procedural guidance and resources for reuse, but a relevant Skill does not necessarily improve task performance. Existing studies

agentsarxiv-cs-lg

Attention via Black-Box Vector Search

arXiv:2610.10135v1 Announce Type: cross Abstract: Sparse attention mechanisms estimate attention over n tokens using a small subset of keys. Many existing approaches use maximum inner product search (

researcharxiv-cs-lg

Batch Augmentation with Unimodal Fine-tuning for Multimodal Fusion of Large Language Models

arXiv:2505.06592v2 Announce Type: replace Abstract: In this paper, we propose batch augmentation with unimodal fine-tuning for multimodal learning. We start with pre-trained unimodal models. We fine-t

researcharxiv-cs-cv

Bookkeeping, Composition, or Unreachable Gold? Reading MemoryAgentBench's Conflict-Resolution Scores Against a Frozen Last-Write Resolver

model-releasesarxiv-cs-cl

arXiv:2610.09193v1 Announce Type: cross Abstract: MemoryAgentBench's Conflict Resolution split is read as measuring 'selective forgetting'. We execute the benchmark's own rule - the newest statement a

Building innovative enterprise agentic AI systems on Vultr at the Agent Arena Hackathon

Vultr demonstrated its composable cloud platform at the Agent Arena Hackathon in San Francisco by partnering with Cerebral Valley to build production‑ready enterprise agentic AI systems using Serverle

agentsvultr

CATune: Structural Constraint-Aware Bayesian Optimization for DBMS Configuration Tuning

arXiv:2610.09276v1 Announce Type: cross Abstract: Modern DBMSs expose hundreds of configuration knobs, resulting in a high-dimensional and heterogeneous search space that makes automated tuning costly

researcharxiv-cs-lg

CIRRA: Dual-Level Continual Instruction Reconciliation with Ongoing Execution for Embodied Robot Agents in Interactive Household Tasks

model-releasesarxiv-cs-cv

arXiv:2610.08862v1 Announce Type: cross Abstract: Household robots must accommodate new user instructions while executing ongoing tasks. Existing agents often regenerate or extensively revise the rema

CoTrace: Data Recipes for Training Terminal Agents with Harness-Model Co-Evolution

arXiv:2610.10426v1 Announce Type: new Abstract: Terminal-agent capability depends jointly on model weights and the runtime harness that formats prompts, binds tools, and handles error recovery. Existi

Dialect-Robust Speech Language Models with Synthetic Pseudo-Dialect Augmentation

arXiv:2610.09321v1 Announce Type: new Abstract: Speech Language Model (SLM) performance often degrades on dialects due to data scarcity. Conventional text-to-speech (TTS) augmentation struggles to cov

FedSSMCoOp: SSM Encoders for light-weight Federated Prompt Learning for Few-shot Classification

arXiv:2610.09907v1 Announce Type: new Abstract: Vision-Language Models (VLMs) have shown strong performance across a wide range of downstream vision tasks, thanks to the complementary information cont

safetyarxiv-cs-cv

GenAI and Agentic AI Exploit Roundup Q3 2026

agentsowasp-genai-security

Coverage period: July 1, 2026 through September 30, 2026 Overview This roundup consolidates selected major AI-related security incidents and exploit disclosures reported during the coverage period. It

Generative AI for Autonomous Driving: Frontiers and Opportunities

arXiv:2505.08854v2 Announce Type: replace Abstract: Generative Artificial Intelligence (GenAI) constitutes a transformative technological wave that reconfigures industries through its unparalleled cap

safetyarxiv-cs-cv

GPU - Vulkan llama.cpp benchmarks sorted by price to performance

model-releasesr-localllama

This table to help anyone looking to build a budget Data Center homelab. I copied the bulk of value based, mid level, decent speed results GPUs and feed it to AI or SI and here are the recommended res

Interesting approach to solving the problem of getting your grokbot to talk to other agents

Interesting approach to solving the problem of getting your grokbot to talk to other agents Connecting agents together safely and having them talk to each other or your friends agents used to suck. Po

agentsemad-mostaque--x

Involution Without Export Is Wasted Effort

A week in Beijing and Shanghai with the people building AI in China some quotes from the article: ByteDance/Doubao is the only major lab keeping its flagship closed, because it is the only one that al

local-air-localllama

IVG-UAV: An Intelligent Voice-Guided UAV System for Autonomous Ripe Fruit Harvesting with Vision-Based Classification and Adaptive Path Planning

arXiv:2610.09398v1 Announce Type: new Abstract: In tropical regions, their agricultural sectors remain highly dependent on manual labor for fruit harvesting. On large-scale farms, this dependency ofte

agentsarxiv-cs-ro

Jevjitsu: Or, How We Tried Generalized Classifiers on Everything

Classification is one of the oldest problems in machine learning. Unlike modern decoding transformers, most classification models were non-autoregressive by nature. Before Jev, the common approaches w

model-releasesqdrant

Kernel Autoresearch for Open-Ended Model Discovery

arXiv:2610.10394v1 Announce Type: new Abstract: Kernels encode the inductive bias of a wide range of machine learning models, yet automated kernel design faces a fundamental dilemma. A fixed grammar o

safetyarxiv-cs-lg

Tutorials

Mechanics of Long-Context Hybrid Models Part 1.1: From Hybrid Attention to Hybrid Position

arXiv:2610.10114v1 Announce Type: new Abstract: The architectural design of Large Language Models (LLMs) is shifting from traditional full-attention-only models to hybrid models, which combine differe

tutorialsarxiv-cs-cl

Multi-Objective Aligned Small Language Model Framework for SUD Patient Dialogue Generation

arXiv:2610.09209v1 Announce Type: new Abstract: Substance Use Disorder (SUD) counseling requires patient responses that reflect underlying cognitive states such as beliefs, coping strategies, and read

OrBIT: Structure-Guided Embedding Compression

arXiv:2610.10385v1 Announce Type: new Abstract: Embedding tables are among the largest components of modern language models. Most compression methods fix a coding geometry such as coordinate blocks, l

local-aiarxiv-cs-lg

Prototype-Based Knowledge Guidance for Fine-Grained Structured Radiology Reporting

model-releasesarxiv-cs-cv

arXiv:2603.11938v2 Announce Type: replace-cross Abstract: Structured radiology reporting promises faster, more consistent communication than free text, but automation remains difficult as models must

Quantize by Drift: Label-Free Mixed-Precision Post-Training Quantization for Text Embedders

arXiv:2610.09227v1 Announce Type: cross Abstract: Mixed-precision post-training quantization needs a per-module sensitivity signal; for a text embedder the obvious one -- the retrieval quality a modul

Tutorials

Robotic Ultra-Long-Horizon Manipulation Skills via Human-guided Lifelong Code Generation

arXiv:2509.18597v3 Announce Type: replace Abstract: Large language models (LLMs) can translate natural-language instructions for robotic manipulation into executable code, but ambiguity, noisy generat

tutorialsarxiv-cs-ro

SAPD: Step-Aligned Privileged Distillation

arXiv:2610.09665v1 Announce Type: new Abstract: On-policy post-training can improve large language models by learning from their own trajectories, but requires costly rollout generation. We ask whethe

Steering Follows Geometry, Not Labels: Emotion Directions in a Full-Duplex Speech Model

arXiv:2610.08887v1 Announce Type: new Abstract: Full-duplex voice agents need to modulate emotion and delivery during real-time conversations, when de-escalating a complaint, carrying urgency in dispa

Thank you :) Swift Models hit 2.2 million+ downloads / Early Access to New Models, Free Compute for Researchers

model-releasesr-localllama

Hey everyone, Jovan from UkisAI (Swift Qwen) here! For those who don't know us, UkisAI is a small lab making tiny frontier LLMs, tools and datasets (+doing it open-source!). I'm one of the guys runnin

Validity Without Ground Truth: What Stated-Preference Economics Offers the Evaluation of Language Models

arXiv:2610.10506v1 Announce Type: cross Abstract: Many of the questions now put to large language models have no correct answer to score against: what a policy is worth, which option a user should cho

Welcome to Gemini at Work 2026: Introducing the Gemini agent

model-releasesgoogle-cloud-ai

Editor’s note: This article is adapted from Thomas Kurian’s keynote address at Gemini at Work 2026. It includes a number of announcements: Gemini, a single, universal agent for work that answers your

A Systematic Study of Small Language Models on Abstract Reasoning Tasks

model-releasesarxiv-cs-cl

arXiv:2610.08680v1 Announce Type: cross Abstract: Endpoint accuracy on abstract-reasoning benchmarks does not reveal whether a language model has acquired a transferable rule or fit distribution-speci

Adaptive Model Inversion Attacks Generalize a Privacy-Robustness Tradeoff

arXiv:2610.07677v1 Announce Type: new Abstract: In this paper, we show that standard evaluations of high-resolution Model Inversion Attacks (MIAs) significantly underestimate training-data privacy lea

researcharxiv-cs-lg

Hardware

AEGIS: Runtime-Guided GPU Collocation for Multi-Tenant Deep Learning Training

arXiv:2508.19073v4 Announce Type: replace-cross Abstract: Deep learning training commonly runs on shared multi-tenant GPU servers, where exclusive allocation provides isolation but can leave resources

hardwarearxiv-cs-lg

Built an SQLite memory engine for Ollama that does not eat all your VRAM

Wanted to share a local memory engine I built called Hillock. The main problem I had with standard local RAG was that running document parsing and vector search chewed up so much VRAM that my actual O

local-air-ollama

Can Semantic Geometry Teach an AI Judgement?

arXiv:2610.07249v1 Announce Type: new Abstract: How can an AI agent determine what rules to follow? One rule permits an action. Another imposes a condition, exception, or conflicting obligation. Deter

safetyarxiv-cs-ai

Catching Developers in the Flow: Low-Latency Agentic Program Repair at Google Scale

arXiv:2610.07289v1 Announce Type: cross Abstract: Manual repair of program failures is time-consuming and disruptive for software developers, particularly during the pre-submit phase where test failur

agentsarxiv-cs-ai

Applications

Cleave: Scaling Tensor Program Optimization via Decoupled Algebraic Search and Operator Scheduling

arXiv:2610.07742v1 Announce Type: cross Abstract: Optimized kernels such as FlashAttention and FlashDecoding are crucial for accelerating today's large models. Most of them are handwritten by experts

applicationsarxiv-cs-ai