Knowledge catalogue
The Keyword: Google introduces DiffusionGemma, an experimental 26B-parameter open model that uses text diffusion for faster text generation compared to autoregressive models — Our newest open experime
arXiv:2601.23231v2 Announce Type: replace-cross Abstract: Flow-based generative models provide strong unconditional priors for inverse problems, but guiding their dynamics for conditional generation r
The post highlights strong market demand for efficient small language models under 9 billion parameters, citing Qwen's top-downloaded models as evidence of user preference for compact, resource-effici
Two bits here I think we're going to see a lot more: 1. custom harnesses / finetunes on smaller open source models to beat frontier models at specific skill or task 2. using frontier models as critics
arXiv:2605.29585v1 Announce Type: new Abstract: Vision-language models (VLMs) are increasingly used to answer questions about physical scenes, yet most evaluations reduce performance to a final answer
arXiv:2605.28769v1 Announce Type: new Abstract: Softmax attention is the cornerstone of modern large language models, but its memory scales linearly and compute quadratically with sequence length. Lin
arXiv:2512.15605v4 Announce Type: replace Abstract: Autoregressive models (ARMs) currently constitute the dominant paradigm for large language models (LLMs). Energy-based models (EBMs) represent anoth
Closed and gated surveillance economy SaaS systems and models will get replaced by open source in two tiers: - The models themselves - The everything-ultra-app harness If an American open source champ
arXiv:2605.22356v1 Announce Type: new Abstract: Large language models are increasingly used as computational tools for modeling human-like behavior. We introduce a behavioral induction framework that
arXiv:2602.16813v3 Announce Type: replace Abstract: Language models based on discrete diffusion have attracted widespread interest for their potential to provide faster generation than autoregressive
arXiv:2602.18895v2 Announce Type: replace-cross Abstract: Large language models (LLMs) have shown promise in translating model-based explanations into human-readable narratives. This study evaluates w
arXiv:2605.16713v1 Announce Type: cross Abstract: Modern Vision-Language Models (VLMs) achieve strong semantic recognition, yet remain brittle on elementary spatial relations such as left of, on, behi
arXiv:2601.23068v2 Announce Type: replace-cross Abstract: Computing the importance of features in supervised classification tasks is critical for model interpretability. Shapley values are a widely us
A new set of open-weight models is topping the leaderboard for document understanding 🔥 INF just released two models: Infinity-Parser2-Pro (35B) and Infinity-Parser2-Flash (2B) that top our @huggingfa
arXiv:2605.11165v1 Announce Type: new Abstract: Federated learning (FL) in heterogeneous environments remains challenging because client models often differ in both architecture and data distribution.
AI Security Institute: Mythos Preview is the first AI model to complete both of AISI's cyber ranges, which measure models' cyberattack capabilities; GPT-5.5 solved only one of them — In February 2026,
arXiv:2605.11167v1 Announce Type: new Abstract: Existing multi-model and tool-augmented systems communicate by generating text, serializing every exchange through the output vocabulary. Can two pretra
arXiv:2603.14186v4 Announce Type: replace Abstract: State-of-the-art text-to-image models produce high-quality images, but inference remains expensive as generation requires several sequential ODE or
Multi-Token Prediction (MTP) for LLaMA.cpp! Running Gemma4 local model 1.5x faster. We patched LLaMA.cpp. Quantized Gemma 4 assistant models into GGUF format. We ran tests on a MacBook Pro M5Max. Gemm
arXiv:2508.15202v2 Announce Type: replace Abstract: Process Reward Models (PRMs) supervise intermediate reasoning steps in large language models (LLMs), but existing PRMs are mainly trained on general
arXiv:2512.05525v2 Announce Type: replace-cross Abstract: Businesses increasingly rely on large language models (LLMs) to automate simple repetitive tasks instead of developing custom machine learning
NIST: An evaluation by NIST's CAISI says DeepSeek V4 Pro lags behind leading US AI models by about eight months and is the most capable Chinese AI model to date — In April 2026, the Center for AI Stan
What people call 'distillation' is a super common practice (you use other models to benchmark your model, to evaluate your inputs or to add a little bit to your datasets) that in my opinion should be
Evelyn Cheng / CNBC: Alibaba says its Qwen AI models will be integrated into BYD, Volkswagen, and other cars, letting users buy food and tickets via voice commands on select models — BEIJING — Chinese
🚨BREAKING: Hugging Face just open-sourced an AI intern that reads ML papers, trains models, and ships the final model for you. It’s called ML Intern. And this is not another AI coding demo that prints
Bloomberg: DeepSeek releases its new flagship models V4 Pro and V4 Flash in preview, saying V4 Pro trails the performance of state-of-the-art models by about 3 to 6 months — DeepSeek rolled out previe
arXiv:2604.01965v2 Announce Type: replace-cross Abstract: Scientific knowledge discovery increasingly relies on large language models, yet many existing scholarly assistants depend on proprietary syst
arXiv:2604.19598v1 Announce Type: cross Abstract: This study compared repeated generation consistency of exercise prescription outputs across three large language models (LLMs), specifically GPT-4.1,
APIs and limited releases for AI models are not a safety policy, they’re a business model (which is totally ok as long as you’re transparent about it). Especially on cyber-security, they give a false
It’s time to go beyond language models. Introducing Odyssey-2 Max, our most powerful world model yet. It materially advances the SOTA in physical accuracy. This is a big step toward models that simula
arXiv:2509.13047v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) have demonstrated remarkable capabilities across many domains, yet their application to specialized fields remain
great read on why open harnesses have become so important. we used to focus on open models vs closed models, but i think the deeper issue is open harnesses vs closed harnesses. if the harness is close
arXiv:2610.15343v1 Announce Type: new Abstract: World models and video understanding models are increasingly deployed in dynamic visual environments, yet their robustness to physical adversarial threa
arXiv:2606.20310v2 Announce Type: replace Abstract: Video reward models are central to aligning video diffusion models with human preferences, yet existing ones are external vision-language models tha
Quartermaster (my local inference manager: one server that swaps llama.cpp, image, audio and video models in and out of VRAM) can now help you server and manage decision models: Auto-discovered. Drop
arXiv:2603.16148v2 Announce Type: replace Abstract: Spiking neural networks promise brain-inspired neural dynamics, yet native large-scale spiking language-model pretraining remains largely unexplored
arXiv:2607.16204v2 Announce Type: replace Abstract: Recent growth in reinforcement learning (RL) has surfaced a need for diverse, specialized training environments. Hand-curated environments with fixe
Preamble and disclaimer, Sample size: 8, at any size of form this research is just screwing around being writing it down. This was just a little,(well... big), curiosity test I wanted to run to see wh
arXiv:2510.19773v2 Announce Type: replace Abstract: Membership inference attacks (MIAs) have emerged as the standard tool for evaluating the privacy risks of AI models. However, state-of-the-art attac
arXiv:2610.09134v1 Announce Type: new Abstract: Robot world models support policy evaluation, planning, and synthetic data generation, but these applications require predictions that distinguish succe
arXiv:2610.07599v1 Announce Type: cross Abstract: In this paper, we study robust decision-making in the latent space of world models (WMs). Robust optimization is a mathematical framework where, given
arXiv:2610.05051v1 Announce Type: new Abstract: Time-series data are often sampled irregularly at high frequencies and exhibit long-range dependencies, which makes long-horizon modelling difficult. Co
arXiv:2610.02267v1 Announce Type: new Abstract: Agent harnesses make many small, typed decisions per task: which model to call, which tool to use, whether retrieved text is relevant, whether an input
Very good to see more competent labs training open-weight models. This Reflection's new Beam model seems to have strong reasoning efficiency, which I think is a big deal for long-running agents. 3-4x
arXiv:2610.02066v1 Announce Type: new Abstract: As Large Language Models (LLMs) increasingly serve as foundational reasoning engines, their tendency to hallucinate remains a critical vulnerability. Wh
arXiv:2609.36245v1 Announce Type: new Abstract: Latent reward models (LRMs) enable efficient alignment of video diffusion models by scoring intermediate states directly in latent space. However, we fi
arXiv:2609.37989v1 Announce Type: new Abstract: Tabular foundation models achieve strong zero-shot accuracy on structured data by pretraining on synthetic tables, but they ignore the column names, tas
arXiv:2609.34414v1 Announce Type: new Abstract: Predicting the next state is a core paradigm of World Models for modeling physical dynamics, emphasizing prediction fidelity. As World Models evolve int
arXiv:2604.05904v2 Announce Type: replace-cross Abstract: Gray-box RC models are widely used to enable energy-efficient model predictive control (MPC) in buildings. However, estimating RC parameters r
I maintain a starter kit where the LLM provider is a runtime-swappable DB row, not a hardcoded constant — OpenAI, Anthropic, or a local Ollama instance. Mostly so I can demo without burning API credit
arXiv:2609.27408v1 Announce Type: cross Abstract: Benchmarks for vision-language models offer their answer choices in some convention: a letter, a color name, a pixel coordinate. That convention is tr
arXiv:2609.25053v1 Announce Type: new Abstract: Can one language model hand its live memory to another without the receiver rereading the context? We demonstrate useful persistent hybrid-state transfe
Worth a read if you track open weights. Nathan Lambert (interconnects) posted the written version of his Congressional testimony on where open models stand. The numbers that got me: Hugging Face downl
arXiv:2609.18950v1 Announce Type: new Abstract: A robot's learned model of its own dynamics is only valid until those dynamics change: actuators wear, payloads shift, and joints stiffen. A model-based
arXiv:2609.15070v1 Announce Type: new Abstract: Diffusion Language Models (DLMs) generate text by iteratively denoising a masked sequence, independently predicting multiple tokens at each step. This c
arXiv:2609.03247v1 Announce Type: cross Abstract: Large language model (LLM) security has largely focused on role-playing jailbreaks, with less attention to what happens when a user asks an LLM to ver
arXiv:2609.06912v1 Announce Type: new Abstract: Tabular foundation models (TFMs) are commonly pretrained on large collections of procedurally generated synthetic tasks, yet it remains unclear how well
arXiv:2609.05968v1 Announce Type: new Abstract: Aligning a text-to-image generation flow model with a reward makes it follow objectives that the training data alone does not provide. Alignment fine-tu
there is a difference between 'i am ok with models being trained on my data and released publicly later' and 'i am ok with internal models trained on my very recent efforts, being used to race the sol
arXiv:2609.02885v1 Announce Type: new Abstract: Recent web agents use world models for test-time action selection by sampling candidate actions, predicting the resulting web states, and ranking them w