All entries

Metadata Predictability Is Not Evidence Dependence: An Intervention-Based Audit for Weak-Label Benchmarks

arXiv:2605.23701v1 Announce Type: new Abstract: We study a protocol-level test for weak-label benchmarks: whether benchmark outputs change when the provided evidence is intervened on. Metadata-only sh

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Industry

Microsoft economist's hot take: Let it burn first

DGX agent

This post discusses how OpenAI is burning 17 billion annually despite 20+ billion in revenue because unit economics for AI inference are fundamentally broken, with the company losing more money on que

industryr-chatgpt

25 May 2026

Millimeter-wave Imaging for Anthropometric Body Measurement

arXiv:2605.23064v1 Announce Type: new Abstract: Body shape and circumferences are clinically informative biomarkers for risk stratification, including measures such as waist to hip ratio, limb and tru

safetyarxiv-cs-cv

Hardware

MiniCPM5-1B is an impressive release in the 1B class! @OpenBMB https://huggingface.co/collections/openbmb/minicpm5 ✨ 1B - Apache 2.0 ✨ Hybri…

MiniCPM5-1B is an impressive release in the 1B class! @OpenBMB https://huggingface.co/collections/openbmb/minicpm5 ✨ 1B - Apache 2.0 ✨ Hybrid reasoning with Think / No-Think modes ✨ 128K context ✨ Run

hardwareclem-delangue--x

MirrorCheck: Efficient Adversarial Defense for Vision-Language Models

arXiv:2406.09250v3 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) are increasingly susceptible to sophisticated adversarial attacks, including adaptive strategies specifically de

Mitigating Object Hallucinations via Sentence-Level Early Intervention

arXiv:2507.12455v3 Announce Type: replace Abstract: Multimodal large language models (MLLMs) have revolutionized cross-modal understanding but continue to struggle with hallucinations - fabricated con

researcharxiv-cs-cv

Model Collapse as Cultural Evolution

model-releasesarxiv-cs-ai

arXiv:2605.23054v1 Announce Type: cross Abstract: Model collapse, the progressive degradation of LLMs trained on their own outputs, has been characterized statistically but lacks a linguistic explanat

ModeSwitch-LLM: A Lightweight Phase-Aware Controller for Cross-Mode LLM Inference on a Single GPU

arXiv:2605.23057v1 Announce Type: cross Abstract: ModeSwitch-LLM is a lightweight request-boundary controller for improving single-GPU large language model inference efficiency by routing each request

Moonwalk: Inverse-Forward Differentiation

model-releasesarxiv-cs-ai

arXiv:2402.14212v4 Announce Type: replace-cross Abstract: Backpropagation's main limitation is its need to store intermediate activations (residuals) during the forward pass, which restricts the depth

Tools

More details on the Datasette blog: https://datasette.io/blog/2026/jump-menu/

Datasette, an open-source tool for exploring and publishing data, released a blog post detailing a new jump menu feature for 2026. The jump menu likely provides improved navigation and quick access to

toolssimon-willison--x

Industry

More than 5,500 GitHub repositories were infected with malware in a supply chain attack, dubbed Megalodon, on May 18 that relies on automated commits (Ionut Arghire/SecurityWeek)

Ionut Arghire / SecurityWeek: More than 5,500 GitHub repositories were infected with malware in a supply chain attack, dubbed Megalodon, on May 18 that relies on automated commits — Fake automated com

industrytechmeme

Move on Muon : A Hamiltonian probability gradient flow perspective of Muon optimizer

model-releasesarxiv-cs-lg

arXiv:2605.23871v1 Announce Type: cross Abstract: We develop a gradient flow on the space of probability measures defined on matrix-valued parameters induced by regularized Muon, an analytically smoot

MuellerPT: Decomposition Driven Pretraining for Dense Learning in Mueller Polarimetry

arXiv:2605.23840v1 Announce Type: new Abstract: Mueller matrix imaging provides rich, physically meaningful contrast for biomedical tissue analysis, but supervised learning is hindered by scarce dense

researcharxiv-cs-cv

Multi-Floor Exploration for Ground Robots via an Incremental Reachable Graph and Structural Priors

arXiv:2605.23350v1 Announce Type: new Abstract: Autonomous exploration of multi-floor buildings remains challenging for ground robots because conventional 2D and 2.5D maps cannot represent overlapping

agentsarxiv-cs-ro

Multi-Gate Residuals

arXiv:2605.23259v1 Announce Type: cross Abstract: While Attention Residuals has shown some effectiveness in addressing the widespread issue of unbounded activation growth across deep residual layers,

Multi-SpatialMLLM: Multi-Frame Spatial Understanding with Multi-Modal Large Language Models

arXiv:2505.17015v2 Announce Type: replace-cross Abstract: Multi-modal large language models (MLLMs) have rapidly advanced in visual tasks, yet their spatial understanding remains limited to single ima

Multilingual Knowledge Transfer under Data Constraints via Lexical Interventions

arXiv:2605.23885v1 Announce Type: new Abstract: Cross-lingual knowledge transfer is critical for building high-performing multilingual language models for languages with insufficient training data. Wh

researcharxiv-cs-cl

Multilingual Steering by Design: Multilingual Sparse Autoencoders and Principled Layer Selection

arXiv:2605.23036v1 Announce Type: new Abstract: Sparse autoencoders (SAEs) enable feature-level mechanistic interpretability and activation steering in large language models (LLMs), but SAE-based lang

Multimodal Crystal Flow: Any-to-Any Modality Generation for Unified Crystal Modeling

arXiv:2602.20210v2 Announce Type: replace-cross Abstract: Crystal modeling spans a family of conditional and unconditional generation tasks, including crystal structure prediction (CSP) and de novo ge

Multimodal Distribution Matching for Vision-Language Dataset Distillation

arXiv:2605.23482v1 Announce Type: cross Abstract: Dataset distillation compresses large training sets into compact synthetic datasets while preserving downstream performance. As modern systems increas

safetyarxiv-cs-ai

MUSEKG: A Knowledge Graph Over Museum Collections

arXiv:2511.16014v2 Announce Type: replace Abstract: Digitisation in the cultural heritage sector has produced large but fragmented repositories of museum collection data, spanning structured catalogue

Musk hates Altman because Altman deceived him. Altman hates Musk because Musk is an egomaniac. LeCun (who is an equally big egomaniac) hates…

Musk hates Altman because Altman deceived him. Altman hates Musk because Musk is an egomaniac. LeCun (who is an equally big egomaniac) hates Musk because Musk is a jerk. Musk hates LeCun because LeCun

safetygary-marcus--x

Industry

My favorite prompt: a) make a plan for <task> b) orchestrate and launch sub-agents to execute the plan c) validate the results from the sub-…

My favorite prompt: a) make a plan for <task> b) orchestrate and launch sub-agents to execute the plan c) validate the results from the sub-agents d) repeat b and c until you finish the plan Grok Buil

industryelon-musk--x

Naturalistic measure of social norms alignment

arXiv:2605.23420v1 Announce Type: new Abstract: Social norms reflect shared expectations on acceptable behavior. Measuring social norms alignment remains challenging, with existing approaches typicall

safetyarxiv-cs-cl

Near-Optimal Private Linear Regression via Iterative Hessian Mixing

arXiv:2601.07545v2 Announce Type: replace Abstract: We study differentially private ordinary least squares (DP-OLS) with bounded data (X,Y) via sketching-based mechanisms. While Gaussian sketching app

researcharxiv-cs-lg

NeuralBoneReg: An Instance-Specific Label-Free Point Cloud-Based Method for Multi-Modal Bone Surface Registration

arXiv:2511.14286v2 Announce Type: replace Abstract: In computer- and robot-assisted orthopedic surgery (CAOS), patient-specific surgical plans derived from preoperative imaging define target locations

safetyarxiv-cs-cv

NeuroNL2LTL: A Neurosymbolic Framework for Natural Language Translation of Linear Temporal Logic

arXiv:2605.22874v1 Announce Type: new Abstract: Effectively translating between natural language (NL) and formal logics like Linear Temporal Logic (LTL) requires expertise that limits formal verificat

safetyarxiv-cs-ai

NeuroWeaver: An Autonomous Evolutionary Agent for Exploring the Programmatic Space of EEG Analysis Pipelines

arXiv:2602.13473v2 Announce Type: replace Abstract: Although foundation models have demonstrated remarkable success in general domains, the application of these models to electroencephalography (EEG)

agentsarxiv-cs-ai

New paper from Microsoft on Self-Evolving Agent Skills

New paper from Microsoft on Self-Evolving Agent Skills New research from Microsoft Research I see a lot of AI engineers handwriting agent skill docs and hope they generalize. Probably not optimal. Thi

agentsdair-ai--x

Next-Latent Prediction Transformers Learn Compact World Models

arXiv:2511.05963v2 Announce Type: replace Abstract: Transformers replace recurrence with a memory that grows with sequence length and self-attention that enables ad-hoc lookups over past tokens. Conse

safetyarxiv-cs-lg

nice write up from the HuggingFace folks aggregating works on defining agents, harnesses, environments, RL, etc. The more we can roughly hav…

nice write up from the HuggingFace folks aggregating works on defining agents, harnesses, environments, RL, etc. The more we can roughly have a shared vocabulary the better…I still find it confusing (

agentsharrison-chase--x

NLG Evaluation: Past, Present, Future

arXiv:2605.23715v1 Announce Type: new Abstract: Natural Language Generation (NLG) evaluation has changed dramatically since 1990, and will continue to evolve in the future. In 1990, when NLG had close

safetyarxiv-cs-cl

No Coin Left Behind: Maximizing Strategic Surplus Against No-Regret Dynamics

arXiv:2604.05129v2 Announce Type: replace-cross Abstract: We investigate the strategic surplus obtainable against a Follow-the-Regularized-Leader (FTRL) learner with constant step size eta in nimes m

researcharxiv-cs-lg

Non-normal spectral signatures of instability in neural network training dynamics

model-releasesarxiv-cs-lg

arXiv:2605.23476v1 Announce Type: new Abstract: Training instabilities in deep networks - loss spikes, oscillatory convergence, and gradient pathologies - are empirically prevalent but lack a rigorous

Tutorials

Nonlinear Transformations Against Unlearnable Datasets

arXiv:2406.02883v2 Announce Type: replace Abstract: Automated scraping stands out as a common method for collecting data in deep learning models without the authorization of data owners. Recent studie

tutorialsarxiv-cs-lg

Not Too Generative, Not Too Discriminative: The Human Alignment Sweet Spot

arXiv:2605.23819v1 Announce Type: cross Abstract: A central question in computational vision is whether human-like visual representations are better explained by discriminative or generative learning.

safetyarxiv-cs-ai

Tools

Notes on Pope Leo XIV's encyclical on AI

Dropped this morning by the Vatican: Magnifica Humanitas of His Holiness Pope Leo XIV on Safeguarding the Human Person in the Time of Artificial Intelligence. This is a very interesting document. It's

toolssimon-willison

NP-LoRA: Null Space Projection for Subject-Style LoRA Fusion

model-releasesarxiv-cs-cv

arXiv:2511.11051v3 Announce Type: replace Abstract: Low-Rank Adaptation (LoRA) fusion enables the composition of subject and style representations for controllable generation without retraining. Howev

Local Ai

ObjectCache: Layerwise Object-Storage Retrieval for KV Cache Reuse

arXiv:2605.22850v1 Announce Type: cross Abstract: Prefix KV caching has become a key mechanism in LLM serving: it reduces time to first token (TTFT) by avoiding redundant computation across requests t

local-aiarxiv-cs-ai

Occlusion-Aware Physics-Semantic Keyframe Selection for Robust Video Editing

arXiv:2605.23192v1 Announce Type: new Abstract: Video editing has recently achieved remarkable progress with diffusion-based generative models, enabling diverse object-level manipulations from natural

researcharxiv-cs-cv

Local Ai

Ollama com 3x de 3060 12vram em uma placa Machinist ,Xeon com 32gb de memória ram,

This Reddit post discusses a technical setup for running Ollama with three NVIDIA RTX 3060 GPUs (each with 12GB VRAM) on a Machinist motherboard paired with a Xeon processor and 32GB of system RAM. Th

local-air-ollama

On the Infinite Width and Depth Limits of Predictive Coding Networks

arXiv:2602.07697v2 Announce Type: replace-cross Abstract: Predictive coding (PC) is a biologically plausible alternative to standard backpropagation (BP) that minimises an energy function with respect

On the Koopman-Based Generalization Bounds for Multi-Task Deep Learning

arXiv:2512.19199v2 Announce Type: replace-cross Abstract: The paper establishes generalization bounds for multitask deep neural networks using operator-theoretic techniques. The authors propose a tigh

Tutorials

On the Provable Importance of Gradients for Language-Assisted Image Clustering

arXiv:2510.16335v4 Announce Type: replace Abstract: This paper investigates the recently emerged problem of Language-assisted Image Clustering (LaIC), where textual semantics are leveraged to improve

tutorialsarxiv-cs-cv

On the Stability of Spherical Hellinger-Kantorovich Flows and Their Implications for Differential Privacy

arXiv:2605.23879v1 Announce Type: cross Abstract: Gradient-flow sampling interprets a Gibbs distribution as the minimizer of an energy functional over probability measures and generates dynamics conve

researcharxiv-cs-lg

on this, @DavidSacks and I agree.

Gary Marcus and David Sacks found common ground on an unspecified topic, as indicated by Marcus's post on X (formerly Twitter). Without access to the full thread context, the specific point of agreeme

safetygary-marcus--x

One-Forcing: Towards Stable One-Step Autoregressive Video Generation

arXiv:2605.23458v1 Announce Type: cross Abstract: Recent advances have substantially improved real-time interactive video generation in the autoregressive regime. However, most existing few-step autor

One Policy, Infinite NPCs: Persona-Traceable Shared RL Policies for Scalable Game Agents

model-releasesarxiv-cs-ai

arXiv:2605.23652v1 Announce Type: new Abstract: On a 300-persona life-simulation benchmark, pcsp achieves compositional zero-shot persona identification up to 17x above chance, Spearman rho approx 0.7