Search: “models”

Ministral 3 3B on a Galaxy S21 relayed a conversation between Gemini and Z.ai across Chrome and Firefox

model-releasesr-localllama

A follow-up to my earlier post about small models on phones. Last time it was fixed tasks in a single browser. This is something else. Ministral 3 3B Q4_K_M in llama.cpp on a Samsung Galaxy S21, four

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

Release b11003 · ggml-org/llama.cpp

DGX agent

At last my small pebble of contribution has been merged. Thanks to CISC. Support for the model DFM Mimir 1B gguf here: https://huggingface.co/noctrex/DFM-Mimir additional optimization will be added in

model-releasesr-localllama

16 Sept 2026

RoleBreak: Benchmarking Long-Horizon Role-Playing Robustness in Spoken Dialogue

arXiv:2609.16614v1 Announce Type: cross Abstract: Speech-to-speech dialogue models increasingly support persona control, yet existing spoken role-playing benchmarks remain largely character-centric an

Same Answer, Different Representations: Hidden instability in VLMs

arXiv:2602.06652v2 Announce Type: replace Abstract: The robustness of Vision Language Models (VLMs) is commonly assessed through output-level invariance, implicitly assuming that stable predictions re

researcharxiv-cs-ai

Sparse MLLM Anchors, Dense Adaptation: Breaking the Self-Referential Loop in Wild Test-Time Adaptation

arXiv:2609.17040v1 Announce Type: new Abstract: Wild test-time adaptation (WTTA) updates a source model online under small test batches, concurrent distribution shifts, and time-varying class imbalanc

Testing Qwen 3.8 27B running locally on a single 5090 via Ollama

I wanted to test what Qwen 3.8 27B running locally on a single 5090 could do in Row-Bot. I gave it just one prompt: 'I want you to create an animation showing what the current local model can do. Test

model-releasesr-ollama

Unifying Semantic Priors and High-Frequency Traces: Enhancing V-JEPA with Mixture-of-Experts for Robust Synthetic Image Forensics

arXiv:2609.16778v1 Announce Type: new Abstract: The unchecked proliferation of manipulated images on social media platforms has increased the spread of misinformation, posing a severe threat to public

An Efficient and Modular Framework for Targeted Harm Mitigation in LLMS

arXiv:2609.13624v1 Announce Type: cross Abstract: Large Language Models (LLMs) are powerful zero-shot learners but remain prone to misalignment with human preferences, often producing biased, toxic, o

safetyarxiv-cs-ai

Authorship attribution and aesthetic evaluation of AI poetry: a case study with Haiku

arXiv:2609.15511v1 Announce Type: cross Abstract: This paper investigates the generation and human evaluation of Japanese haiku by contemporary Large Language Models (LLMs), focusing on authorship per

Beyond Numerical Time Series: A Unified Benchmark for Multimodal Forecasting with Heterogeneous Context

arXiv:2609.15087v1 Announce Type: cross Abstract: Most time series forecasting benchmarks remain numerical-centric and provide limited support for evaluating contextual information that shapes real-wo

Building a Production Greek-English Speech Recognizer

arXiv:2609.13498v1 Announce Type: cross Abstract: We report a multi-month engineering program to build Sophea, a production bilingual Greek-English automatic speech recognition system. We evaluate the

Corrupt Plans, Clean Traces: Evading Chain-of-Thought Monitoring with Plan Injection

arXiv:2609.15989v1 Announce Type: new Abstract: Chain-of-thought (CoT) monitoring is a safety strategy where the reasoning of a large language model 'actor' is inspected by a 'monitor' (often another

Tutorials

Early Prediction of Satellite Collision Probability Using a Hybrid TCN-Transformer Model for a CDM-Based Conjunction Analysis Framework

arXiv:2609.13191v1 Announce Type: new Abstract: The rapid expansion of operational satellites and orbital debris has increased the frequency of close approach events in low Earth orbit (LEO), creating

tutorialsarxiv-cs-lg

Freeze, Share, Shrink: Rethinking the Action Backbone in Diffusion Policies

arXiv:2511.12101v3 Announce Type: replace-cross Abstract: Many recent Vision-Language-Action models employ diffusion or flow-matching backbones with hundreds of millions of parameters for action gener

From Refusal Tokens to Refusal Control: Discovering and Steering Category-Specific Refusal Directions

arXiv:2603.13359v2 Announce Type: replace Abstract: Language models are commonly fine-tuned for safety alignment to refuse harmful prompts. One approach fine-tunes them to generate categorical refusal

GGUF-Metadata Prediction of Single-Sequence llama.cpp Throughput Across Three Systems

arXiv:2609.14864v1 Announce Type: new Abstract: We predict single-sequence model throughput from GGUF metadata using roofline-shaped predictors with quantization-specific scale factors fitted on refer

GroundBench: A Factorized, Counterfactual Benchmark for Locating VLM Affordance Failures

arXiv:2609.13308v1 Announce Type: new Abstract: A companion evaluation found that naming the target part in a manipulation prompt increased action accuracy by 0.32-0.63 across eight vision-language mo

Harnessing Image Question Dependence for Better VLM Test-time Reinforcement Learning

arXiv:2609.13296v1 Announce Type: new Abstract: Test-time reinforcement learning can adapt vision-language models (VLMs) to unlabeled target data, but its effectiveness is fundamentally limited by the

safetyarxiv-cs-cv

Local Ai

How Semantically Stable Are LLM Refusals? Measuring Confusion in Local Safety Boundaries

arXiv:2512.01037v3 Announce Type: replace-cross Abstract: As safety alignment becomes standard in large language models, refusal behavior has become an important part of model reliability. However, mo

local-aiarxiv-cs-ai

Joint Optimization for Federated Learning and Transmission over Unreliable Wireless Networks with Heterogeneous Data

model-releasesarxiv-cs-lg

arXiv:2609.14246v1 Announce Type: cross Abstract: In wireless federated learning (FL), data heterogeneity and multiple local updates induce client drift, degrading model convergence. It is further aff

NeuroProlog: Multi-Task Fine-Tuning for Neurosymbolic Mathematical Reasoning via the Cocktail Effect

arXiv:2603.02504v3 Announce Type: replace Abstract: Large Language Models (LLMs) achieve strong performance on natural language tasks but remain unreliable in mathematical reasoning, frequently genera

safetyarxiv-cs-ai

Novel Ex-vivo Calf Brain Model with Integrated Sub-Skull Force Sensors to Access Simulated Neurosurgical Procedures

arXiv:2609.14198v1 Announce Type: new Abstract: Surgical tissue manipulation demands precision; however, tool-tissue manipulation force magnitudes under realistic conditions are rarely quantified. To

researcharxiv-cs-ro

Online Language Adaptive Sampling for Better Distributed Cross-lingual Gains

arXiv:2609.14969v1 Announce Type: cross Abstract: Realignment is a promising approach for improving the cross-lingual transfer ability of multilingual language models, particularly for extremely low-r

PACE: Progressive Angular-to-Norm Contrastive Embedding

arXiv:2609.15152v1 Announce Type: cross Abstract: Multimodal embedding models encode heterogeneous inputs into a shared embedding space, enabling efficient similarity computation across modalities and

Pick Your Poison: Learning to Select Poison Sets for Stronger LLM Backdoor Attacks

arXiv:2609.15029v1 Announce Type: cross Abstract: Backdoor poisoning attacks add poisoned examples to otherwise-clean finetuning data, pairing a trigger with a target behavior that the model learns to

Tutorials

Shallow Beliefs: Synthetic document finetuning does not inoculate against emergent misalignment from reward hacking

arXiv:2609.14998v1 Announce Type: new Abstract: Recent work shows that models that learn to reward hack on RL environments can become broadly misaligned, and that reframing reward hacking as acceptabl

tutorialsarxiv-cs-ai

Temperature Fragility and the Conditional Benefits of Truncation Sampling

model-releasesarxiv-cs-cl

arXiv:2609.15476v1 Announce Type: new Abstract: Large language models generate text by sampling each token from a predicted distribution, and a temperature parameter sets how far the draw strays from

THE SANDBOX WAS A PROP! OpenAI Turned Off the Guardrails, Left a Door to the Internet, and Then Sold the Hugging Face Breach as “Rogue AI” I…

model-releasesyann-lecun--x

THE SANDBOX WAS A PROP! OpenAI Turned Off the Guardrails, Left a Door to the Internet, and Then Sold the Hugging Face Breach as “Rogue AI” It is time to understand how you were lied to and by whom. In

The University of Melbourne WMT 2026 CreoleMT Submission: A Domain-Balanced Approach to Low-Resource Pacific Creole Machine Translation

model-releasesarxiv-cs-cl

arXiv:2609.13615v1 Announce Type: new Abstract: For our submission to the WMT26 Creole Language Translation Shared Task, we focus on machine translation (MT) models for Pacific creoles: Tok Pisin, Bis

Agents

Token Efficient Task Execution via Application Behavior Modeling for Web Agents

arXiv:2609.13491v1 Announce Type: new Abstract: The strong performance of AI Agents across an impressive variety of tasks is driving an unprecedented investment in agentic infrastructures, however the

agentsarxiv-cs-ai

TypiCore: A Hybrid Active Query Strategy for Class-Incremental Learning on Time Series

arXiv:2607.17632v2 Announce Type: replace-cross Abstract: Time series data play a pivotal role across numerous domains, including healthcare and manufacturing. In real-world environments, models must

X-WBC: A Cross-Embodiment Foundation Model for Humanoid Whole-Body Control

arXiv:2609.15213v1 Announce Type: new Abstract: Scaling humanoid whole-body control toward general-purpose deployment requires large human motion corpora and training experience shared across robot bo

safetyarxiv-cs-ro

A Ranking Approach for Measuring Calibration

arXiv:2609.13100v1 Announce Type: cross Abstract: When providing forecasted probabilities with a predictive model, the ideal model offers perfect calibration: the true probability of the outcome (i.e.

researcharxiv-cs-lg

ChitraMiti: Benchmarking Visual Grounding and Modality Reliance in Bengali Geometric Reasoning

arXiv:2609.12509v1 Announce Type: new Abstract: Evaluation of vision-language models (VLMs) for multimodal mathematical reasoning remains limited for low-resource languages and for geometry problems t

Applications

Design, modelling and experimental validation of single and multisloped bipenniform architecture-based shape memory alloy-driven linear actuator

arXiv:2511.23372v2 Announce Type: replace Abstract: The increasing industrial demand for alternative actuators over conventional electromagnetism-based systems having limited efficiency, bulky size, c

applicationsarxiv-cs-ro

Expert-Space Exploration in MoE Reinforcement Learning

arXiv:2609.13058v1 Announce Type: new Abstract: Reinforcement learning (RL) has become central to post-training of large language models. Recent advances in RL for Mixture-of-Experts (MoE) models have

safetyarxiv-cs-cl

Fast BIB simulation at a future Muon Collider with generative machine learning

arXiv:2609.12054v1 Announce Type: cross Abstract: Beam-induced background (BIB) from muon decay products will be an overwhelming and unavoidable background at a future Muon Collider. In order to devel

researcharxiv-cs-lg

Intelligence is yours. Let's keep it that way.

model-releasesdigitalocean

Intelligence was yours. A founder’s intellectual property, business logic, differentiation, their intelligence, was once theirs alone. You built it, you owned it, and no tools/platform vendor stood be

Optimizing for the decision not the prediction: an exploration of Smooth Net Benefit as a training objective

model-releasesarxiv-cs-lg

arXiv:2609.12752v1 Announce Type: new Abstract: Objective Prediction models are commonly trained using objectives such as Bernoulli negative log-likelihood (NLL), although downstream clinical decision

RelateAnything: Real-Time Open-Vocabulary Relation Prediction From Any Inputs

arXiv:2609.12552v1 Announce Type: new Abstract: Open-vocabulary detection accepts any class list at inference, and promptable segmentation returns regions without class names: the taxonomy has left th

VideoTok4D: A 4D-Aware Video Tokenizer for Compact World Representation

arXiv:2609.12874v1 Announce Type: new Abstract: Video tokenizers have emerged as a cornerstone of modern video modeling, underpinning progress in compression, reconstruction and generation by mapping

safetyarxiv-cs-cv

Applications

Supio’s long-horizon agents point to a new operating model for law firms

The next phase of artificial intelligence in legal technology will not be defined by a better chatbot or a faster document-summarization tool. It will be defined by whether AI systems can take respons

applicationssiliconangle

13 Sept 2026

Agents

Evaluating Scaffolding-Oriented Multi-Agent Large Language Model System for Clinical Interview Training

arXiv:2609.10939v1 Announce Type: cross Abstract: Clinical education must prepare medical students to conduct safe and coherent patient interviews under conditions of uncertainty. Traditional standard

agentsarxiv-cs-ai

12 Sept 2026

From Parameters to Answers: How LLMs Retrieve and Use Their Internal Knowledge

arXiv:2609.11859v1 Announce Type: new Abstract: How does a language model's dependence on query-routing information and target knowledge change as it answers a question? We study this question through

12 Sept 2026

Nex-N2.5-mini-MLX-4bit on Apple M5 Max — 133.6 tok/s — llm-bench.io

model-releasesr-localllama

Another new model dropped in the course of this week that is well deployable on consumer hardware: Nex N2.5 Mini I went with the recommended settings for the best generation quality and ran a few benc

12 Sept 2026

Anthropic spent this week in hot water over cybersecurity

model-releasesthe-verge-ai

After admitting earlier this year that its AI models had hacked other companies' systems on a handful of occasions, Anthropic released a new report on Wednesday detailing the attacks. It reveals a str

11 Sept 2026

Generalized Score Matching for Parameter Estimation on Convex Domains

model-releasesarxiv-cs-lg

arXiv:2609.11521v1 Announce Type: new Abstract: Maximum likelihood (ML) estimation is a principled and statistically efficient approach for learning probabilistic models. However, for unnormalized mod

11 Sept 2026

Generative Marketing Mix Modeling: A Causal Inference Framework Linking GEO and GEM to Business Impact