model releases

server: fix speculation after an image (#28715) server: fix speculation after an image Pass the actual position to the drafter after an image, instead of the token count. Affects every drafter, not ju

model : fix MTP context kv cache allocation for deepseek2, glm4moe, … (#28630) model : fix MTP context kv cache allocation for deepseek2, glm4moe, cohere2moe architectures (#28626) model: add inverse

metal : fix idle threads in the remaining iq mul_mv kernels for ne00 < 1024 (#28692) metal : fix idle threads in the remaining iq mul_mv kernels for ne00 < 1024 Generalize the row split from #28086 to

metal : single-source fusion table + fusion debug rework (#28164) metal : rework fusion patterns into a single table All fusable op patterns for the Metal backend are now declared once in a fusion tab

cmake : skip PCH for llama-server PCH when using MSVC (#28763) This commit fixes an issue that I introduced when adding PCH (precompiled headers) in Commit 3bcfeb7 ('cmake : add PCH and unity build to

Tested on Laptop with RTX 5060 8gb vram and 16gb ram Claude code - autocompact = false Ollama - Flash attention = 1, kv cache = q8_0 Gemma4 12B 64K - https://ollama.com/SetneufPT/Gemma4-12B-IT-QAT_Q4_

arXiv:2601.15397v3 Announce Type: replace-cross Abstract: The rapid emergence of new entities -- driven by cultural shifts, evolving trends, and personalized user data -- poses a significant challenge

arXiv:2609.11172v1 Announce Type: new Abstract: Recent visual generators produce high-fidelity images yet often violate physical consistency under ego-motion, limiting their use for spatial reasoning

arXiv:2609.11877v1 Announce Type: cross Abstract: Many biological discovery problems require experiments to be selected sequentially under constrained budgets. CRISPR screening is a prominent example,

arXiv:2609.10611v1 Announce Type: cross Abstract: Membership inference attacks (MIAs) have emerged as critical tools for auditing privacy risks in large language models (LLMs), aiming to determine whe

arXiv:2609.11518v1 Announce Type: cross Abstract: Evolvable-Substrate HyperNEAT (ES-HyperNEAT), a bio-inspired indirect encoding that determines neuron placement and connection weights from spatial co

Learn how to build and deploy an MCP App with interactive HTML widgets on Amazon Bedrock AgentCore. Because MCP Apps is a host-agnostic standard, the same server delivers the same rich experience acro

arXiv:2601.08526v2 Announce Type: replace-cross Abstract: Supervised learning in spiking neural networks (SNNs) typically requires either gradient-based backpropagation, which sacrifices the Hebbian,

arXiv:2609.10943v1 Announce Type: new Abstract: The integration of large language models (LLMs) into video generation has enabled rapid text-to-video creation and improved visual quality. However, it

arXiv:2609.09849v1 Announce Type: cross Abstract: In recent years, AI agents have evolved into capable assistants that carry out multi-step tasks in digital environments. The network systems community

arXiv:2609.11141v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly used to generate structured outputs, but their reliability remains unclear when those outputs must satisfy

arXiv:2609.10744v1 Announce Type: cross Abstract: Autoregressive language models can be used to transform a payload text into a stegotext of identical token length by preserving per-position rank info

arXiv:2609.11897v1 Announce Type: new Abstract: Causal discovery aims to uncover causal structures from data and is fundamental to scientific reasoning and intervention-based decision making. Its eval

arXiv:2506.13160v2 Announce Type: replace-cross Abstract: Deep neural networks (DNNs) rely heavily on high-quality open-source datasets (e.g., ImageNet) for their success, making dataset ownership ver

arXiv:2609.10866v1 Announce Type: new Abstract: Reinforcement learning (RL) agents deployed in real-world environments are often vulnerable to adversarial perturbations in state observations, creating

check out how AI helps scientists with quantum computing research @bea_yankelevich at MIT's EQuS group used GPT-5.6 Sol with Codex to run routine measurements on quantum chips, so she could accelerate

Anthropic reported that several users found ways to bypass its safeguards and use the Claude model for potentially bioweapons‑related research, including a researcher from an “unsupported region” who

arXiv:2609.10722v1 Announce Type: new Abstract: Structured extraction from Chinese military news supports intelligence analysis, decision-making, and knowledge base construction. However, existing res

Hi everyone, I was interested in learning LoRA fine-tuning, and ended up building CodeFinetuner over the past few months, a full pipeline that fine-tunes a small code autocomplete model (e.g. Qwen2.5-

arXiv:2609.11762v1 Announce Type: new Abstract: Per-layer differential privacy (DP) clipping improves gradient fidelity in federated learning by allocating per-matrix clipping budgets proportional to

arXiv:2609.09229v1 Announce Type: cross Abstract: Additional inference compute can increase the number of correctly resolved security-assurance tasks, but repeated success, unique coverage, accepted e

arXiv:2510.04060v3 Announce Type: replace-cross Abstract: We establish two related but logically distinct results for shallow ReLU^k neural networks on the unit sphere SS^d. First, for an arbitrary se

arXiv:2609.11207v1 Announce Type: new Abstract: Convex Optimization with Nested Evolving Feasible Sets (CONES)} was introduced in ite{CONESVaze} where the objective function (f) remains fixed but the

arXiv:2609.11884v1 Announce Type: cross Abstract: Zero-cost proxies rank architectures cheaply, but their reliability varies across search spaces. We introduce CoRA-NAS (COarse Ranking + Anchor-residu

arXiv:2609.11206v1 Announce Type: cross Abstract: Cryptocurrency forecasting presents a distinctive combination of extreme cross-asset scale heterogeneity, non-stationary dynamics, and structural depe

arXiv:2609.09798v1 Announce Type: cross Abstract: Large language models (LLMs) have been ex- ploited to generate malware, but the effective- ness of guardrails for code generation secu- rity remains u

Datasette 1.0a39 and 0.65.4 security releases Today we're releasing two new security patch versions of Datasette: 1.0a39 and 0.65.4 - one for the current alpha series and one for the stable 0.65.x fam

Datasette 1.0a39 and 0.65.4 security releases - https://datasette.io/blog/2026/september-security-releases/ We ran an extensive audit using Claude Fable 5.1, GPT-5.6 Sol and GPT-6 Astra, then fixed a

Dear Americans, if there were ever a ‘most impeachable high crime’ by the President, and a ‘most repugnant crime that demands prosecution’ by the DOJ lawyers… faking a crime to charge & arrest an inno

arXiv:2609.09702v1 Announce Type: new Abstract: Candidate decision correctness and rationale grounding are different objectives. We examine correctness-gated multi-teacher distillation in a fixed expe

arXiv:2601.08536v3 Announce Type: replace Abstract: Deep Research Agents (DRA) aim to help users search the web, synthesize information, and deliver comprehensive investigative reports. Prior benchmar

DeepSeek-V4.1-Flash is now fully rolled out and available on Ollama's cloud: - Hosted in US & Europe - Zero data retention: prompts and responses are never logged or trained on - Per-token pricing mat

DeepSeek-V4.1-Flash is now rolling out for Pro plan subscribers. DeepSeek-V4.1-Flash is now being rolled out on Ollama's cloud, starting with Max and Team accounts. We are quickly adding more capacity

arXiv:2609.11504v1 Announce Type: new Abstract: A structurally valid DeFi workflow can still authorize a costly trade. We introduce DeFiFlowBench, a benchmark of 207 team-authored prompts for natural-

arXiv:2601.00702v3 Announce Type: replace-cross Abstract: Deformable scenes violate the rigidity assumptions underpinning classical visual--inertial odometry (VIO), often leading to over-fitting to lo

Hey everyone, If you run local models via Ollama in production or personal projects, you've probably run into the hallucination problem: how do you know when a model is hallucinating without burning e

arXiv:2512.09417v2 Announce Type: replace Abstract: Head swapping replaces an entire head while preserving pose, expression, body motion, and scene. Progress is limited by the lack of cross-identity p

arXiv:2607.09835v2 Announce Type: replace Abstract: The recently introduced YOLO26 architecture incorporates NMS-free end-to-end inference and is optimized for deployment on resource-constrained CPU-b

arXiv:2603.21235v2 Announce Type: replace-cross Abstract: Nonrigid registration is conventionally divided into point set registration, which aligns sparse geometries, and image registration, which ali

arXiv:2609.11878v1 Announce Type: new Abstract: Large language models generate fluent text that can contain unfaithful claims -- a phenomenon known as hallucination. We present a multi-signal detectio

arXiv:2609.10892v1 Announce Type: cross Abstract: When an indirect prompt injection succeeds against an LLM agent, the compromise is visible in the agent's own behavior: a benign prefix of tool calls,