model releases

Stop Probing, Start Coding: Why Linear Probes and Sparse Autoencoders Fail at Compositional Generalisation

arXiv:2603.28744v2 Announce Type: replace Abstract: The linear representation hypothesis states that neural network activations encode high-level concepts as linear mixtures. However, under superposit

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

StoreBench: A Live-Commerce Environment for Evaluating and Training Autonomous Operator Agents

DGX agent

arXiv:2610.10942v1 Announce Type: new Abstract: Reinforcement learning environments are now a primary lever for improving large language model (LLM) capabilities in post-training, yet most agentic ben

model-releasesarxiv-cs-ai

9 Oct 2026

Strategic Governance of AI Models in Earth Science

arXiv:2610.10560v1 Announce Type: cross Abstract: AI foundation models pretrained on weather and climate data are increasingly fine-tuned to Earth science tasks well beyond weather forecasting. Their

Strategic Investment Decision Making for Value Creation in Energy Transition: A Reinforcement Learning Approach

arXiv:2610.10768v1 Announce Type: cross Abstract: The global challenge of climate change has driven significant steps to reduce CO2 emissions, guided by international agreements like the Paris Agreeme

Strix halo + CMP 170HX setup

model-releasesr-localllama

A bit of backstory : i got lucky in August and was able to grab a CMP 170 HX 8GB for around 500 on alibaba. I tried later to get a second one for around 1k but my order got cancelled and prices went t

SV-TAD: Native Sparse Convs for Efficient Temporal Action Detection

arXiv:2610.11579v1 Announce Type: new Abstract: To adapt billion-parameter Vision Transformers for long-video understanding, recent methods freeze the backbone and train lightweight convolutional modu

SWE-Journey: Towards More Realistic Evaluation of Coding Assistants through Long-Horizon, Multi-Turn Interaction

arXiv:2610.11559v1 Announce Type: cross Abstract: Coding assistants such as Claude Code and Codex have become a major application of LLM agents, yet existing benchmarks remain far from real-world use,

Symbolic Density Estimators for Unnormalized Distributions

arXiv:2610.10807v1 Announce Type: new Abstract: Estimating the symbolic or analytical form of probability density functions (PDFs) from observed samples is a fundamental challenge in statistical and c

Syn-Omni: Structured Specialization and Progressive Collaboration for Omnimodal Embeddings

arXiv:2610.12256v1 Announce Type: cross Abstract: Omnimodal embeddings naturally involve both shared representations and modality-specific features across heterogeneous inputs. However, existing omnim

Tell Robot What Not to Do: A Negation Understanding Perspective

model-releasesarxiv-cs-ro

arXiv:2610.11952v1 Announce Type: new Abstract: Instruction following enables robots to perform diverse tasks specified in natural language, making it a fundamental capability for human-robot interact

tencent/Youtu-Parsing-Omni · Hugging Face

model-releasesr-localllama

Youtu-Parsing-Omni is a compact (5B) omni-modal parsing model. Given a single input — a document page, a natural image, a chart / flowchart, a geometry figure, an audio clip or an audio-visual video —

Test-Time Compute for Tabular Foundation Models: Mechanisms, Gains, and Limits

arXiv:2610.12005v1 Announce Type: new Abstract: Which forms of test-time compute improve the predictions of strong pretrained tabular foundation models (TFMs)? We systematically study this along three

Tested Mellum2.1-12B-A2.5B on PI Coding Agent - surprisingly usable, but not great at one-shot projects

model-releasesr-localllama

I tested Mellum2.1-12B-A2.5B (Q8) locally using Pi and llama-server. All five tests were one-shot. Results were pretty mixed: - Pelican SVG, Browser OS, Minecraft: Poor results. - Bouncing Hexagon: Ph

TestJack: Should you trust the results in coding benchmarks? Agentic Coding Benchmarks Auditing via Evaluator Evolution

arXiv:2610.10619v1 Announce Type: cross Abstract: Large language model (LLM) agents are rapidly reshaping software engineering, accompanied by an explosion of new code benchmarks. Yet nearly all exist

The challenge for Google post-Gemini 4 will be what they do with a good model. Anthropic & OpenAI show we are moving towards a single interf…

model-releasesethan-mollick--x

The challenge for Google post-Gemini 4 will be what they do with a good model. Anthropic & OpenAI show we are moving towards a single interface for many tasks with orchestrator agents. The Gemini 3 er

ThinkingBox: Solving an agent task once vs. solving it 20/20: 507 stateful workflows graded on terminal database state [R]

model-releasesr-machinelearning

Figure 1b from our paper Disclosure: I'm one of the authors (Microsoft). The paper, code, dataset are public and ThinkingBox is on Hugging Face OpenEnv as well. Raw evaluation trajectories are not rel

TKCAM: Text and Keyframe to Camera Trajectory Generation

arXiv:2610.11105v1 Announce Type: new Abstract: Generating high-quality and controllable camera motion is essential for AI-assisted cinematography, video synthesis, and 3D scene understanding. We intr

TokenRouter: Efficient Serving System for Token-Level LLM Routing

arXiv:2610.12242v1 Announce Type: new Abstract: Large language model (LLM) routing distributes inference work across different models, advancing the cost-quality Pareto frontier of LLM serving. While

Towards Unified Evaluation of Prompt Enhancers for Video Generation

arXiv:2610.11736v1 Announce Type: new Abstract: Modern video generators can realize increasingly complex visual narratives, positioning the prompt enhancer (PE) as a critical bridge from concise user

TRACE: Diagnosing Verifier Brittleness in Agentic Evaluation

arXiv:2610.11678v1 Announce Type: new Abstract: Verifier scores now serve as both benchmark metrics and training rewards for large language model (LLM) agents, and a change in score is routinely read

Training on the Future: A Delay-Aware Audit of Test-Time Adaptation for Time-Series Forecasting

arXiv:2610.12232v1 Announce Type: new Abstract: Test-time adaptation (TTA) methods for time-series forecasting update a deployed model, or a small adapter around it, from incoming ground truth. But th

TRL v1.15 is out, and it’s an absolute banger of a release for memory-efficient post-training. The main change: SFT, DPO, KTO, GRPO, RLOO an…

model-releasesclem-delangue--x

TRL v1.15 is out, and it’s an absolute banger of a release for memory-efficient post-training. The main change: SFT, DPO, KTO, GRPO, RLOO and Distillation now use a fused LM head by default. Instead o

ttok 1.0

model-releasessimon-willison

Release: ttok 1.0 I released ttok 0.4, ran uv tool upgrade ttok, piped a file into the new version... and realized that it was defaulting to the GPT-4 tokenizer when it should very clearly default to

TypedBench: A Benchmark for Calibration, Framing Sensitivity, and Cost in System One Decision Models

arXiv:2610.11392v1 Announce Type: new Abstract: System One models output calibrated probabilities over typed answers such as categorical choices, ordinal levels, or binary outcomes, via a non-generati

UltraLight Luma: A Novel Edge-Deployable Perception Network for Crop-Row Segmentation in Agricultural Robotics

model-releasesarxiv-cs-ro

arXiv:2610.11764v1 Announce Type: new Abstract: Reliable crop-row perception is essential for autonomous agricultural robots, but low-cost deployment is constrained by computation, memory, power, and

Uncovering and Fixing Collider Bias in Bayesian PINNs

arXiv:2610.11737v1 Announce Type: cross Abstract: Bayesian physics-informed neural networks (B-PINNs) are a popular framework for parameter and state inference from sparse or noisy observations. They

Universal Textual Teaching for LLMs

arXiv:2610.12114v1 Announce Type: new Abstract: Knowledge distillation (KD) transfers knowledge from stronger Teacher models to weaker Student models, but most methods require training the Student par

Use and Disuse: Intent-Structured Experience Consolidation for Memory and Learning in LLM Agents

arXiv:2610.12124v1 Announce Type: new Abstract: The evolution of Large Language Model agents from single-task execution to long-term autonomous operation highlights the critical challenge of transform

Using LMs to Model the Effects of Context and Coreference during Sentence Comprehension

arXiv:2609.32119v2 Announce Type: replace Abstract: Language models (LMs) are often used as a tool to model human language processing. Recent studies suggest that severely restricting LMs' context win

V-ECE: Estimating General Expected Calibration Errors

arXiv:2602.24230v2 Announce Type: replace-cross Abstract: In probabilistic classification, calibration error (CE) measures the average divergence of predicted probabilities f(X) from P(Y|f(X)), the tr

VibeEdit: Image Editing with Canvas Instructions

arXiv:2610.12229v1 Announce Type: cross Abstract: In text-guided image editing, describing the desired change is often straightforward, but identifying the intended object or region can be cumbersome,

VioLA: Learning Generalist Humanoid Control Policies from Human Data

arXiv:2610.12435v1 Announce Type: cross Abstract: Teaching a humanoid to follow instructions with its whole body runs into two obstacles. Its action space is large and tightly coupled: legs, arms, and

Visible Reasoning Is Not a Universal Optimizer: Persona- and Thinking-Dependent Effects in Analytics Code Generation

arXiv:2610.10639v1 Announce Type: cross Abstract: Visible Chain-of-Thought (CoT) is often treated as a broadly useful reasoning instruction, yet analytics code generation combines natural-language amb

WARP-VLA: Wrist-Camera Adaptation for View-Robust Policy Execution in Vision-Language-Action Models

arXiv:2610.11508v1 Announce Type: cross Abstract: Despite recent advances in Vision-Language-Action models (VLAs) for robotic manipulation, their performance remains sensitive to changes in camera con

Web automation with Ollama on a 4 GB GPU: Qwen 3.5 4B plans, a heuristic clicks, and saved tasks replay with no model at all

I've been building ANCHOR, a web automation agent that runs entirely on Ollama. You describe a task in plain language ('Register Maria Silva, IT department, contractor, and save'), point it at a site,

model-releasesr-ollama

What 30,000 Hours of Ego-centric Video Does Not Teach

arXiv:2610.12464v1 Announce Type: new Abstract: World models offer a promising alternative to physics-based simulators, yet remain far from practical deployment. We ask how far scaling ego-centric hum

What to Admit and How to Present: Governing Persistent Memory in LLM Agents

arXiv:2610.11188v1 Announce Type: new Abstract: Persistent memory can improve personalization in LLM agents but can also induce sycophancy and cross-domain leakage. We distinguish two governance decis

When Citations Mislead? A Claim-Level Benchmark for Legal Hallucination Detection

arXiv:2610.10971v1 Announce Type: cross Abstract: Large language models are increasingly used in legal research and drafting, but they can still produce claims that sound convincing without being supp

When History Helps and Hurts: Selective History Use across Multimodal Turns

arXiv:2610.11948v1 Announce Type: new Abstract: Reliable multimodal interaction depends on selective use of conversational history: an earlier question may remain relevant while its previous answer is

When Interfaces Speak: Data-Aware Generative UI Harness for Active Interaction

arXiv:2610.11123v1 Announce Type: new Abstract: Most human-agent interaction today remains text-based. Natural language can impose cognitive overload, ambiguity, information chaos, and slow input for

Where to Adapt Matters: Layer-Selective Fine-Tuning for Capability Retention

arXiv:2610.11620v1 Announce Type: new Abstract: Parameter-efficient fine-tuning (PEFT) enables large language models (LLMs) to adapt to specialized tasks, but often at the cost of degrading general ca

Which Skill to Distill? SGUID: Selecting a Compact Skill Bank for Model-Skill Co-Evolution

arXiv:2610.12367v1 Announce Type: new Abstract: Skills, reusable procedural guidance added at inference, can substantially improve LLM downstream performance (Li et al., 2026). Prior work retrieves sk

Why LLM Agents Favor Their Group: Stakes, Observed Norms, and Reputation

arXiv:2610.11008v1 Announce Type: cross Abstract: Language-model agents favor their own group because they have watched their members favor each other. The group label alone does little once the decis

Why On-Policy Distillation Sometimes Fails: Vanishing Learning Signals

arXiv:2610.11247v1 Announce Type: cross Abstract: On-policy distillation (OPD) enables effective capability transfer between language models, yet the mechanisms underlying its failures are not fully u

Workerville: Towards an Organizational Behavior Account of Agent Safety

arXiv:2610.11561v1 Announce Type: new Abstract: LLM-based agents now interact with their environments continuously, shaped by such organizational channels as user instructions, peer messages, and long

WorldBench: Evaluating LLMs on Three.js Voxel World Generation

arXiv:2610.10622v1 Announce Type: cross Abstract: Large language models can now write complete, interactive 3D worlds as code, but grading those worlds automatically is unreliable. Existing judges tak

WorldFact-Bench: Beyond Image-Internal Plausibility to Image-World Consistency

arXiv:2610.11184v1 Announce Type: new Abstract: Advances in image generation have made visual authenticity increasingly difficult to assess. Although image forensics now examines both generation artif

WOVEN: Weaving Visual World Modeling into Multimodal LLMs

arXiv:2610.12417v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) struggle with spatial, embodied, physical, and temporal reasoning. We hypothesize that these failures reflect