| 85 |
SO-Arena: Benchmarking the Capability Spectrum and Reliability of System One Models · National University of Singapore · code (24★) |
| 49 |
BirdDreams: Joint Perspective and BEV Video Generation for a Plausible, Reactive Driving World Model · NVIDIA |
| 47 |
Scaling Mathematical Proof Supervision from Structured Textbooks · Intern Large Models |
| 45 |
PhysLoop: Harnessing LLMs' Physics Awareness for Self-Improving Video Generation Models · DAGroup-PKU · code (34★) |
| 44 |
G-STAR: Group-Scaled Token Capacity for Deeply Compressed Autoregressive Generation · LongCat · code (25★) |
| 40 |
One Table, Many Queries: Benchmarking Generalist Tabular Models Across Generation, Imputation, and Prediction · Stable AI · code (1★) |
| 39 |
Duo: A Reciprocal World Model for Full-Duplex Interaction · inclusionAI · code (40★) |
| 36 |
SteerDistill: Quantizing LLM Agents via Closed-Loop Teacher Steering · code (10★) |
| 32 |
Actor, Judge, or Teacher? Understanding Decision Models in Long-Horizon Agents · Zhejiang University · code (42★) · on AI Squawk |
| 32 |
Open Vision Banana: A Fully Open Recipe for Generative Visual Perception · Fundamental AI Lab at UTN · code (21★) |
| 30 |
Incentivizing Self-Learning in Large Language Models via Reinforcement Learning · Intern Large Models |
| 27 |
On the Design of Hyper-Connections: Simplification, Stability, and Scalability · dots studio · code (20★) |
| 26 |
Ropedia Xperience 2: A Gold Standard for Scaling Physical AI Environments and Behaviors · Ropedia · code (16★) |
| 22 |
Read Less, Solve More: Token-Efficient Sparse Reading of External Content for AI Agents · Institute of Computing Technology · code (38★) |
| 22 |
TTB: Test-time Baking for Efficient Rendering of Decoder-only View Synthesis Models · NVIDIA · code (35★) |
| 18 |
Learning the Task or Learning the Verifier? Rethinking Criteria Exposure in Embodied Self-Improvement · DAGroup-PKU |
| 16 |
WorldDraft: If Code Builds the World, Can Video Models Just Render It? · code (10★) |
| 14 |
Unbundling Multimodal Vision Encoding · Meta · on AI Squawk |
| 13 |
Ensemble Routing for Prompt-Adaptive Cost-Aware LLM Inference |
| 13 |
Interactive Visual Gym: Evaluating Visual Interaction Models with Simulated Users in Interactive Worlds · code (8★) |
| 12 |
GTAGym: AAA Game Tasks Environment for Evaluation and Reinforcement Learning · OpenMOSS · code (11★) |
| 12 |
Evolving Infra: Agent-Adaptive LLM Serving at Scale · MIT HAN Lab · code (30★) |
| 12 |
UniSpark: Exploring Hybrid Attention for Multimodal Understanding and Generation · Ant Group |
| 12 |
Factored One-sided Adam-Moment for Practical and Scalable SOAP · NAVER AI Lab · code (0★) |
| 9 |
Test-Time Scaling for Diffusion Language Models with Frequency-Aware Remasking · UC Riverside |
| 6 |
Where Does Alignment Emerge and What Does It Mean? A Closer Look at the Platonic Representation Hypothesis · Technical University of Munich · code (8★) |
| 6 |
FAME-QP: A GPU-Accelerated Quadratic-Programming Solver for Real-Time Model Predictive Control · The Ohio State University · code (11★) |
| 6 |
DeepStress: Stress-Testing Deep Search Agents · Orange |
| 6 |
Who Remembers What? Multi-Granularity Memory Extraction for Multi-User, Multi-Agent Interaction · Amazon Web Services |
| 5 |
Subliminal Learning in Vision-Language Models · Johns Hopkins University |
| 4 |
CogBattery: Human-Referenced Evaluation of Cognitive Task Behavior and Structure in Language Models · code (1★) |
| 4 |
WMIP: A Working-Memory-Inspired Programmable Probe of Cumulative State Tracking in Large Language Models · code (1★) |
| 4 |
Precision on Demand: Escalating Weight Bits at Uncertain Decoding Steps in a Small Local Model |
| 4 |
Agent-R^2: Scaling Agents for Realistic Productivity Work via Real-Source Data Synthesis · Zhejiang University · code (7★) |
| 4 |
JustRUA: A Controlled Study of GPT-6 Robot-Use Agents · code (1★) |
| 4 |
TurtLES: A Large-Scale Benchmark for Turbulent 3D Neural PDE Surrogates · Sorbonne University, Institute of Intelligent Systems and Robotics · code (1★) |
| 3 |
Prompt2Echo (P2Echo): Text-Prompted Multi-View Echocardiography Segmentation with One Model · University of British Columbia · code (5★) |
| 3 |
Visual Similarity in Adult Word Extension and Artificial Word Learning |
| 3 |
Module-Aware Optimization for Graph Neural Networks · code (3★) |
| 3 |
Mixed Signals: Weight-Space Programming for EEG and ECG · EPFL |
| 3 |
CARTBench: Rubric-Based Evaluation of Shopping Advice for Complex and Ambiguous Requests with Timelines · code (0★) |
| 3 |
Probing Transformer Representations Layer by Layer, Module by Module · code (1★) |
| 3 |
Evidence-Guided Recovery for AI Agents with Ambiguous Tool Outcomes and Partial Execution |
| 3 |
Do Multimodal Models Decide Consistently on Equivalent Optical Spectra? · code (1★) |
| 3 |
Humanized but Still True? Auditing Facts and Detection in AI-Text Humanizers · code (1★) |
| 3 |
HiVe-Mem: Interleaving Horizontal Memory Expansion and Vertical Evidence Composition for Multimodal Agents · Nanyang Technological University · code (0★) |
| 3 |
Experience-Handoff Reinforcement Learning for Recursive Self-Improvement · KAIST AI · code (3★) |
| 3 |
Consensus Capture in Reference-Based UAV Video Stabilization: A Benchmark on Dynamic Urban Scenes · EPFL · code (0★) |
| 3 |
Back to Pixels: Lifting Latent Video Diffusion into Pixel Space · Roblox Corporation |
| 2 |
Bits or permutations? Spending a fixed inference budget on weight precision or on option-order averaging in small local language models |
| 2 |
PolyRoute: A Twelve-Language and Code-Mixed Benchmark for Typed-Decision Routing, with Baselines and a Leaderboard |
| 2 |
ScrapeBench: Evaluating Legal Compliance of AI Agents in Website Scraping · code (0★) |
| 2 |
From Large Language Models to Efficient System One Models via Adaptive Computation · code (1★) |
| 2 |
Structured Confidence Perturbation via Fisher Geometry for Adversarial Training · XJTLU |
| 2 |
A Cross-Scenario Survey and Analysis of Self-Evolving Multimodal Agents · code (0★) |
| 2 |
The Narrative Gap: Can LLMs Help Us Navigate Diverse Narratives Across Languages? · Johns Hopkins University · code (0★) |
| 2 |
The Value of Thinking: Learning When to Reason in Hybrid Decision Models · code (1★) |
| 1 |
LO-World: Unified BEV Latent Modeling for Joint 4D Occupancy and LiDAR Forecasting · code (5★) |
| 1 |
A frontier machine intelligence for human neuroimaging · Machine Learning in Neurosurgery Lab · code (4★) |
| 1 |
EgoX: Egocentric Cross-Embodiment Manipulation with Embodiment Dreaming · Carnegie Mellon University · code (9★) |