Search: “models”

I just published my roadmap for building evals on highly subjective tasks. Here are the key steps involved and practical tricks for reducing…

I just published my roadmap for building evals on highly subjective tasks. Here are the key steps involved and practical tricks for reducing noise / increasing confidence… Blog is here: https://camero

model-releasesswyx--x

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

I use local AI to generate personal value on my massive NAS-based archive / library

DGX agent

I’ve seen vastly more questions than answers about how people are using their local AI so I figured I’d make a post about my experience. It seems most people are looking for frontier-level coding expe

model-releasesr-localllama

21 Oct 2026

Learning the 4D World through Point Trajectory Forecasting

arXiv:2610.28594v1 Announce Type: new Abstract: Reliable multi-view forecasting requires videos from different cameras to describe the same evolving 3D scene. To this end, we introduce Video-Point Mod

LiBRA: Logit Fingerprints for Byzantine-Robust Client Selection in Federated Learning

arXiv:2610.27357v1 Announce Type: cross Abstract: Byzantine-robust aggregation shields federated learning from malicious updates, but the rules that filter out suspect updates do so by comparing clien

Look Before You Reason: Implicit Visual Thinking for Efficient Multimodal Reasoning

arXiv:2610.27341v1 Announce Type: new Abstract: Vision-language models (VLMs) have advanced rapidly in multimodal reasoning, yet they remain limited on tasks requiring fine-grained understanding of sm

LoWAgent: A Low-Resource Multilingual Benchmark for Web Agents

arXiv:2610.27205v1 Announce Type: cross Abstract: Web agents are typically evaluated on English-language interfaces, leaving their ability to operate across the world's multilingual websites largely u

LSA-PAST: Layer-Specific Adaptation and Prototype Transport for Compositional Incremental Learning

arXiv:2610.27203v1 Announce Type: new Abstract: Class-Incremental Learning (CIL) enables models to recognize new classes over time while retaining prior knowledge, but many real-world problems are com

MindView: Benchmarking Generative Multi-View Spatial Intelligence

arXiv:2610.28622v1 Announce Type: cross Abstract: Spatial intelligence has recently been assessed through generative tasks such as spatially grounded image editing, yet language-guided generation of s

Local Ai

MoDDiT: Simple Token Routing in Diffusion Transformers

arXiv:2610.28086v1 Announce Type: cross Abstract: Real-time video generation, interactive world models, and on-device image synthesis are almost a reality. Diffusion Transformers (DiT) achieve excelle

local-aiarxiv-cs-ai

Navalia: Advancing Cowork Agents through Verifiable Task Synthesis and Long-Horizon Post-Training

arXiv:2610.26887v1 Announce Type: new Abstract: Large language model (LLM) agents are evolving from conversational assistants into Cowork Agents capable of autonomous planning, tool use, and complex t

OPTIMA: Optimal One-shot Pruning for LLMs via Quadratic Programming Reconstruction

arXiv:2512.13886v3 Announce Type: replace-cross Abstract: Post-training pruning cuts the inference cost of large language models, yet it faces a trade-off: simple heuristics that zero weights are fast

Push-Pull Attentional Anchoring for Diffusion Concept Erasure

arXiv:2610.27719v1 Announce Type: new Abstract: Rapid advances in diffusion models have raised concerns about privacy, harmful content, and copyright violations. Concept erasure methods aim to address

Qwen researching Qwen

model-releasesr-localllama

I always feel a little bad asking LLMs to do research into itself or newer versions. It's like asking someone to face their own mortality, although that's anthropomorphizing. The user is asking me to

Applications

RAG-NIMBF: A Non-Invasive, Modular Benchmarking Framework for Automated RAG Configuration Analysis

arXiv:2610.27816v1 Announce Type: cross Abstract: Retrieval-Augmented Generation (RAG) has become the dominant paradigm for grounding the outputs of large language models (LLM) in external knowledge.

applicationsarxiv-cs-ai

RealControlBench: A Benchmark for Interpretability-Informed Fine-Grained Control of LLMs and Agents

model-releasesarxiv-cs-cl

arXiv:2610.27295v1 Announce Type: new Abstract: Existing steering benchmarks are built around simplified, isolated generation tasks and evaluate a single behavioral attribute under controlled settings

Recognition Is Not Grounding: Benchmarking Multilingual Financial Evidence Extraction

arXiv:2610.19432v2 Announce Type: replace Abstract: Comparing financial disclosures across languages requires identifying shared accounting concepts and recovering all numerical facts that express the

Agents

Revision-Based Learning from Interactive Experience and Feedback

arXiv:2610.26636v2 Announce Type: replace Abstract: Code generation models are increasingly used to produce interactive artifacts, such as websites and user interfaces. Many properties that determine

agentsarxiv-cs-ai

Tutorials

Sample-efficient learning with Dense Associative Memories

arXiv:2610.27384v1 Announce Type: new Abstract: We analytically study the evolution of the energy function of Dense Associative Memories (DenseAMs) as they train on randomly sampled codewords of rando

tutorialsarxiv-cs-lg

SCODEL: Shared Canonical Ontology Definitions for Food Entity Linking and Ontology-Grounded Recipe Knowledge Graphs

arXiv:2610.27637v1 Announce Type: new Abstract: Identifying recipe entities and linking them to ontology concepts is fundamental for semantic recipe understanding. However, food mentions in recipes ar

Agents

Secure Agentic Coding through Counterexample-Grounded Feedback

arXiv:2610.25916v2 Announce Type: replace-cross Abstract: As coding models and agent harnesses become more capable, reliable security verification is becoming a central bottleneck. Improving secure co

agentsarxiv-cs-ai

Teaching AI Agent to Play Poker Strategically Through Preference Optimization

arXiv:2610.27028v1 Announce Type: new Abstract: A poker agent can make a legal move and still make a strategically poor one. We ask whether compact AI agents can exhibit strategic behavior under hidde

Agents

The Dynamics of Collapse and Recovery in Clustered Federated Learning

arXiv:2610.28565v1 Announce Type: new Abstract: Clustered federated learning (CFL) can suffer from cluster collapse. In particular, under unfavorable initialization, early client assignments concentra

agentsarxiv-cs-lg

Training-Free Continual Learning for Multimodal Forecasting via Prompt Evolution

arXiv:2610.27468v1 Announce Type: cross Abstract: Adapting a forecasting system to a changing world classically means retraining. We ask whether the task of continually learning regimes for time serie

Tried running 748B DeepSeek-V4.1-Flash on a €2k PC

model-releasesr-localllama

I bought a used PC with an RTX 3090 24 GB, 128 GB DDR4 and an Intel i9-9900X last year for under €2,000. With the release of DeepSeek-V4.1-Flash-1017 I decided to try running it on my machine. It is o

When Do LLMs Help with Feature Selection? A Systematic Empirical Study

arXiv:2610.28556v1 Announce Type: cross Abstract: Large language models (LLMs) can leverage semantic feature names alongside summary statistics for tabular feature selection, yet it remains unclear wh

When does Subgroup Balance Matter? Predicting Performance Sensitivity from Latent Separability

arXiv:2512.09496v2 Announce Type: replace-cross Abstract: Unequal representation of demographic groups in training data poses challenges to model generalisation across populations. Standard practice a

researcharxiv-cs-ai

Which Depth Estimator for Jetson AGX Orin? A 2026 Benchmark of Monocular and Stereo Methods

arXiv:2610.27199v1 Announce Type: cross Abstract: The Jetson AGX Orin is a common computer for robots and vehicles, yet measured accuracy, speed and energy of current depth estimators on it are hard t

A Decision-Theoretic View of Test-Time Training: When, How Far, and Which Directions to Adapt

arXiv:2606.15569v2 Announce Type: replace Abstract: Test-time training (TTT) adapts a pretrained model to each prompt via parameter updates, improving accuracy under pretraining-to-test distribution s

An Empirical Study of Parameter-Efficient Fine-Tuning for Mixture-of-Experts LLMs

arXiv:2610.26370v1 Announce Type: new Abstract: Parameter-efficient fine-tuning (PEFT) methods are critical for adapting large models under resource-constrained settings. However, as mixture-of-expert

Best way to accurately annotate 500 images for fine-tuning Meta DETR?[P]

researchr-machinelearning

Hey everyone! I'm working on an object-detection project and fine-tuning Meta's DETR model on a custom dataset of around 500 images. The main challenge I'm facing right now is annotation. I need to la

Beyond Omni: A Unified MLLM for Embodied Sensory Understanding

arXiv:2610.26751v1 Announce Type: new Abstract: Recent embodied systems increasingly build on multimodal large language models (MLLMs) for their pretrained understanding and reasoning capabilities, bu

Tutorials

CasePlay: Self-Play Reinforcement Learning from Case Reports for Medical Reasoning

arXiv:2610.24952v1 Announce Type: new Abstract: Reinforcement learning (RL) is a promising way to improve the clinical reasoning abilities of large language models (LLMs), but current pipelines often

tutorialsarxiv-cs-lg

Celigo launches MCP tool management for enterprise AI assistants and agent access

model-releasessiliconangle

Intelligent automation startup Celigo Inc. today announced its Model Context Protocol Management platform, which will let information technology and security teams govern how artificial intelligence a

Consistent LLM Call Handling Across Java Frameworks: A Reference Implementation and Benchmark

arXiv:2610.25480v1 Announce Type: new Abstract: Java applications reach large language models (LLMs) through an integration framework, and rely on it to handle a failing provider: to retry what can be

Design Anything, Design at Scale

arXiv:2610.24771v1 Announce Type: new Abstract: As multimodal resources are transformed into media outputs, agentic AI is expected to stay aligned with design intent over long horizons. Designers foll

EMMA: Event-Driven Multimodal Memory System with Multi-Agent Retrieval

arXiv:2610.23892v1 Announce Type: new Abstract: Long video understanding requires reasoning over hours of content that far exceed the context window of multimodal large language models (MLLMs). Existi

First-Order LoRA Training with Phone NPUs

arXiv:2610.24816v1 Announce Type: new Abstract: On-device fine-tuning lets a large language model (LLM) learn from its user's private data without uploading it. Low-rank adaptation (LoRA) reduces the

For the next 3 days, in collaboration with @MiniMax_AI, MiniMax M3.1 Flash will be free on Ollama's Cloud! Hosted in the US and with zero da…

For the next 3 days, in collaboration with @MiniMax_AI, MiniMax M3.1 Flash will be free on Ollama's Cloud! Hosted in the US and with zero data retention. MiniMax M3.1 Flash is on Ollama! 🔒 Hosted in t

model-releasesollama--x

From Label Priors to Task Evidence: Efficient Long-Video Frame Selection via Bayesian GRPO

arXiv:2610.25418v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) face significant scalability challenges in long-video understanding due to quadratic attention complexity and c

GliomaINK: A Tumor and Anatomy Conditioned Flow Matching Framework for 3D MRI Sequence Synthesis in Glioma Patients

arXiv:2610.24044v1 Announce Type: cross Abstract: Background: Incomplete multiparametric MR examinations limit glioma characterization and downstream AI tools. Generative models can synthesize missing

GPT-6 Astra Is a Native Visuo-Tactile Policy for Contact-Rich Manipulation

model-releasesarxiv-cs-ro

arXiv:2610.26846v1 Announce Type: new Abstract: Recent advances in multimodal models have enabled GPT-6 Astra to perform robotic manipulation through in-context learning. Yet contact-rich control and

Safety

Hidden Trajectories Don't Lie: Correcting Corruption Drift for Test-Time Adaptation

arXiv:2610.24220v1 Announce Type: new Abstract: Test-time adaptation (TTA) adapts pretrained models to distribution shifts encountered during deployment. However, most existing methods rely on the wei

safetyarxiv-cs-lg

Inside the Loop: A Mechanistic Study of Weight-Tied Transformers on Depth-Bound Algorithmic Tasks

arXiv:2610.24135v1 Announce Type: new Abstract: Solving algorithmic tasks with many sequential steps typically requires scaling up transformer depth or chain-of-thought length, incurring unfavorable p

IntriAdapt: Generalizable Intrinsic-Guided Test-Time Adaptation for Metric Geometry Estimation

arXiv:2610.24607v1 Announce Type: new Abstract: Recovering metric 3D geometry from images is fundamentally ill-posed due to perspective projection ambiguity. Current geometry models incorporate suppli

Learning volatility-adaptive forgetting via GARCH-modulated LSTM for stock price forecasting

arXiv:2610.24386v1 Announce Type: cross Abstract: Stock price forecasting is hindered by the heteroscedastic nature of financial time series, in which turbulent and tranquil regimes alternate. Althoug

Learning When to Abort in Transformer Planning: A Mechanistic Study

arXiv:2610.25300v1 Announce Type: new Abstract: Planning does not by itself imply the ability to recognise errors, but recognising errors can improve planning capacity. In this paper, we study this ph

researcharxiv-cs-lg

LLM-Flock: LLM-Driven Decentralized Multi-Robot Flocking via Influence-Based Consensus with Collision Avoidance

model-releasesarxiv-cs-ro

arXiv:2505.06513v2 Announce Type: replace Abstract: Large Language Models (LLMs) have shown strong capabilities in reasoning and planning, prompting researchers to explore their use as decentralized d

Machine Unlearning in Low-Dimensional Feature Subspace

arXiv:2601.22456v2 Announce Type: replace Abstract: Machine Unlearning (MU) aims at removing the influence of specific data from a pretrained model while preserving performance on the remaining data.

researcharxiv-cs-lg