Search: “model-releases”

arXiv:2609.26618v1 Announce Type: new Abstract: End-to-end (E2E) driving policies have progressed rapidly on open-loop (OL) benchmarks, yet OL evaluation cannot reveal whether a policy withstands comp

New paper! 🫡 We introduce Matryoshka Attribution, a new attribution method which uses gradient descent to find which parts of a neural network are responsible for a behaviour. MAttr is #1 on the Mecha

NextLM Inc., which makes software that uses behavioral analysis and custom artificial intelligence models to help sales professionals identify and rank individual sales leads, is bringing its AI Prosp

arXiv:2609.25022v1 Announce Type: cross Abstract: Line-segment detection is fundamental to robotics, autonomous navigation, and industrial inspection. While transformer-based detectors achieve the hig

arXiv:2609.26402v1 Announce Type: new Abstract: The discovery of novel inorganic materials drives technological breakthroughs in critical fields such as computing and energy storage. Generative AI has

arXiv:2609.23407v1 Announce Type: cross Abstract: Humans can effortlessly localize the direction of a sound source and integrate it with visual cues for reasoning, yet this remains challenging for emb

On stage with @MistralAI . 🔥 From building models to putting them into the hands of developers, a closer look at how Mistral is shaping the next generation of AI products. Straight from the #AIEParis

arXiv:2609.26280v1 Announce Type: new Abstract: In this chapter, we investigate how classical machine learning models respond to small, targeted modifications in their parameters. We compare and contr

I have been using open-webui and open-terminal. Open-terminal is installed in a VM and gives the agent a ton of control of it's own system that I keep away from my normal network. I find this gives th

OpenAI: OpenAI releases MentalHealthBench, an open benchmark to evaluate AI responses in realistic mental health conversations, developed with 80+ licensed experts — An open benchmark developed with m

Zoe Kleinman / BBC: OpenAI says it will give its AI cyber defense system Daybreak and GPT-5.6 Sol to the Ukrainian government for free to help it protect civilian infrastructure — The tech giant OpenA

arXiv:2609.25710v1 Announce Type: cross Abstract: The statistical accuracy of neural networks depends on both their approximation power and the complexity of the class fitted from data. While increasi

arXiv:2507.15836v2 Announce Type: replace Abstract: In this work we study black-box privacy auditing, where the goal is to lower bound the privacy parameter of a differentially private learning algori

arXiv:2609.26052v1 Announce Type: new Abstract: Diffusion Large Language Models (dLLMs) have recently emerged as a promising alternative to conventional Auto-Regressive (AR) Large Language Models (LLM

Out today: Nemotron 3 Diarization, so you can implement speaker detection in your voice applications. 🗣️✅ And it's available day-0 on transformers and @Dell Enterprise Hub! Enterprises can implement o

arXiv:2609.22770v2 Announce Type: cross Abstract: We study a parameterized hybrid ranker that fuses a dense embedding list and a sparse lexical list. The method has a small, explicit parameter vector:

arXiv:2609.25396v1 Announce Type: new Abstract: Parallel coding agents can produce patches that work alone but fail when merged. This happens when one agent changes an interface or rule that another a

arXiv:2609.25046v1 Announce Type: new Abstract: Peer review plays a central role in scholarly publishing, yet verifying whether reviewer claims are supported by manuscript evidence remains a largely m

arXiv:2609.25381v1 Announce Type: cross Abstract: We propose penalized nonreversible Langevin algorithms for sampling from pi(x)propto e^{-f(x)}mathbf 1_{mathcal C}(x), where mathcal Csubsetmathbb R^d

I am proud to release these quants of Qwen 3.8 27B. They beat the excellent ISTA and Unsloth quants byte-for-byte on three corpora. Both KLD and top 1% were tested 3x. It took a week of continuous GPU

arXiv:2609.26795v1 Announce Type: new Abstract: 3D Gaussian Splatting (3DGS) can reconstruct a captured scene photorealistically, but the resulting representation does not by itself support physical i

arXiv:2609.23695v1 Announce Type: new Abstract: Recent advances in Physical AI have accelerated the use of foundation models in autonomous systems such as unmanned aerial vehicles (UAVs), which must p

If like me you enjoyed classics like Baldur's Gate 2, this is a small fun experiment. For anyone interested: https://www.youtube.com/live/8FhPfRKTucw?si=PMIVdHbZ-31PNwZD Qwen 3.8 Flash Next has amazin

arXiv:2609.26310v1 Announce Type: new Abstract: Graph neural networks have achieved strong performance in node classification by aggregating information from graph neighborhoods. However, a single agg

arXiv:2503.06989v5 Announce Type: replace-cross Abstract: Recently, Multimodal Large Language Models (MLLMs) have demonstrated their superior ability in understanding multimodal content. However, they

arXiv:2609.25827v1 Announce Type: new Abstract: Reported gains in vessel-trajectory prediction from Automatic Identification System (AIS) data are credited to new architectures, but the evaluation pro

Thought it's about time to share after testing for a week. You need four things most people miss: the right quant, the right model, the right branch, and the right cache flags. https://github.com/dtm-

arXiv:2609.25176v1 Announce Type: cross Abstract: Real-time voice assistants must reason over evolving requests, execute actions, and follow conversational rules. Qwen-Audio-3.1-Realtime brings these

Got QFN up and running on our Strix box this past weekend and have been running on it for a few days now. Big thank you/shout out to the Halogen team, it's running fantastic on the Strix, this is clea

arXiv:2609.25611v1 Announce Type: new Abstract: We introduce Qwen3.8-Omni-Flash, a natively multimodal agentic model for real-world multimodal productivity. Compared with previous omni models, which p

First: That the 7GB (int8) 2.1 comes even close to the 20GB (int8) 2511 is amazing, but maybe because of the smaller size some capabilities got lost. For some tasks 2.1 it is much better than 2511, bu

Qi2.1 Viggle 0.2, 5steps, CFG2.1 prompt: Cute young woman at beach holding Grumpy Cat. We're still getting a soft focus with the 'baseline'steps of 5, non-baseline CFG 2.1.... Now she is wearing a swe

arXiv:2609.26069v1 Announce Type: cross Abstract: Simulation-based tutor selection can be unstable when predictively adequate learner models imply different policy rankings. RankCert certifies one of

Compare Jev and GPT-5.4 nano for real-time LLM guardrails, with demo results on latency, cost, and checks on agent inputs, replies, and tool calls. The post Real-time LLM guardrails with Jev: comparin

arXiv:2609.25515v1 Announce Type: new Abstract: Standard deployment-ready object detectors for autonomous vehicles degrade in adverse weather and lighting conditions without being trained on extensive

arXiv:2609.24710v1 Announce Type: cross Abstract: Large Language Models (LLMs) are increasingly used in cybersecurity, where accurate analysis often requires multi-step and context-dependent reasoning

arXiv:2609.25028v1 Announce Type: new Abstract: QMSum provides no scorer, making query-focused meeting summarization results difficult to compare. We rescore or generate 15 systems under one implement

arXiv:2609.25602v1 Announce Type: cross Abstract: In language models, the choice between believing the prompt and believing the weights is made by a handful of identifiable attention heads. Instructio

arXiv:2507.21726v3 Announce Type: replace-cross Abstract: Tree tensor networks (TTNs) are widely used in low-rank approximation and quantum many-body simulation. In this work, we present a formal anal

arXiv:2609.25636v1 Announce Type: new Abstract: Modern embodied agents achieve impressive success rates, yet their actual instruction-following ability is far weaker than these numbers suggest. We tra

arXiv:2609.26292v1 Announce Type: new Abstract: Physical-condition diversity is largely missing from current benchmarks for robot manipulation. While large-scale simulation benchmarks increasingly inc

arXiv:2609.25633v1 Announce Type: cross Abstract: Dynamic 3D Gaussian splats (3DGS) model time-varying scenes using a separate Gaussian set per frame. While neighboring video frames are highly correla

Rolling out starting today: — Developers: Both models in @GoogleAIStudio and the Gemini API — Consumers: Gemini 3.8 Flash TTS in @Gemini_Notebook and Gemini 3.8 Flash-Lite TTS in Google Vids — Coming

arXiv:2609.12551v2 Announce Type: replace-cross Abstract: AI is beginning to make substantive contributions to LLM inference optimization. Existing AI optimizations are predominantly profiling-based.

arXiv:2609.25567v1 Announce Type: cross Abstract: A lack of high-throughput phenotyping solutions for root traits in field-grown crops has severely constrained understanding and improvement of below-g

arXiv:2609.25270v1 Announce Type: new Abstract: Generating Scalable Vector Graphics (SVG) code from natural-language instructions is an open-ended task without absolute visual ground truth, leaving bo