Rising Stars in Research

WorldSimBench: Towards Video Generation Models as World Simulators

From Crowdsourced Data to High-quality Benchmarks: Arena-Hard and Benchbuilder Pipeline

SparseVLM: Visual Token Sparsification for Efficient Vision-Language Model Inference

Aguvis: Unified Pure Vision Agents for Autonomous GUI Interaction

Training Software Engineering Agents and Verifiers with SWE-Gym

AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs

Layer by Layer: Uncovering Hidden Representations in Language Models

Imagine While Reasoning in Space: Multimodal Visualization-of-Thought

Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs

Taming Rectified Flow for Inversion and Editing

A General Framework for Inference-time Scaling and Steering of Diffusion Models

Freeze-Omni: A Smart and Low Latency Speech-to-speech Dialogue Model with Frozen LLM

AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders

MedXpertQA: Benchmarking Expert-Level Medical Reasoning and Understanding

OR-Bench: An Over-Refusal Benchmark for Large Language Models

Theoretical guarantees on the best-of-n alignment policy

MME-CoT: Benchmarking Chain-of-Thought in Large Multimodal Models for Reasoning Quality, Robustness, and Efficiency

Learning Smooth and Expressive Interatomic Potentials for Physical Property Prediction

Towards World Simulator: Crafting Physical Commonsense-Based Benchmark for Video Generation

Scaling Test-Time Compute Without Verification or RL is Suboptimal

Cradle: Empowering Foundation Agents towards General Computer Control

History-Guided Video Diffusion

GuardAgent: Safeguard LLM Agents via Knowledge-Enabled Reasoning

SWE-Lancer: Can Frontier LLMs Earn $1 Million from Real-World Freelance Software Engineering?

LMRL Gym: Benchmarks for Multi-Turn Reinforcement Learning with Language Models

HealthGPT: A Medical Large Vision-Language Model for Unifying Comprehension and Generation via Heterogeneous Knowledge Adaptation

ShadowKV: KV Cache in Shadows for High-Throughput Long-Context LLM Inference

RE-Bench: Evaluating Frontier AI R&D Capabilities of Language Model Agents against Human Experts

AutoML-Agent: A Multi-Agent LLM Framework for Full-Pipeline AutoML

Sundial: A Family of Highly Capable Time Series Foundation Models

MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models

The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity

Video-R1: Reinforcing Video Reasoning in MLLMs

Why Do Multi-Agent LLM Systems Fail?

VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning

Training Language Models to Reason Efficiently

ToolRL: Reward is All Tool Learning Needs

Mean Flows for One-step Generative Modeling

Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach

VITA-1.5: Towards GPT-4o Level Real-Time Vision and Speech Interaction

SuperGPQA: Scaling LLM Evaluation across 285 Graduate Disciplines

TTRL: Test-Time Reinforcement Learning

Improving Video Generation with Human Feedback

Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning

ProRL: Prolonged Reinforcement Learning Expands Reasoning Boundaries in Large Language Models

EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test

Show-o2: Improved Native Unified Multimodal Models

Remarkable Robustness of LLMs: Stages of Inference?

WebDancer: Towards Autonomous Information Seeking Agency

RetrievalAttention: Accelerating Long-Context LLM Inference via Vector Retrieval

OWL: Optimized Workforce Learning for General Multi-Agent Assistance in Real-World Task Automation

LMFusion: Adapting Pretrained Language Models for Multimodal Generation

Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Models

The Surprising Effectiveness of Negative Reinforcement in LLM Reasoning

General-Reasoner: Advancing LLM Reasoning Across All Domains

Offline Actor-Critic for Average Reward MDPs

Simplicity Prevails: Rethinking Negative Preference Optimization for LLM Unlearning

Video-RAG: Visually-aligned Retrieval-Augmented Long Video Comprehension

Spatial-MLLM: Boosting MLLM Capabilities in Visual-based Spatial Intelligence

SWE-smith: Scaling Data for Software Engineering Agents

LLaVA-CoT: Let Vision Language Models Reason Step-by-Step

R1-Onevision: Advancing Generalized Multimodal Reasoning through Cross-Modal Formalization

CoTracker3: Simpler and Better Point Tracking by Pseudo-Labelling Real Videos

LVBench: An Extreme Long Video Understanding Benchmark

R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimization

LLaVA-3D: A Simple yet Effective Pathway to Empowering LMMs with 3D Capabilities

DimensionX: Create Any 3D and 4D Scenes from a Single Image with Decoupled Video Diffusion

GUIOdyssey: A Comprehensive Dataset for Cross-App GUI Navigation on Mobile Devices

OmniHuman-1: Rethinking the Scaling-Up of One-Stage Conditioned Human Animation Models

Stable Virtual Camera: Generative View Synthesis with Diffusion Models

REPA-E: Unlocking VAE for End-to-End Tuning of Latent Diffusion Transformers

MV-Adapter: Multi-View Consistent Image Generation Made Easy

EasyControl: Adding Efficient and Flexible Control for Diffusion Transformer

MeshAnything V2: Artist-Created Mesh Generation with Adjacent Mesh Tokenization

GameFactory: Creating New Games with Generative Interactive Videos

ORION: A Holistic End-to-End Autonomous Driving Framework by Vision-Language Instructed Action Generation

DriveArena: A Closed-loop Generative Simulation Platform for Autonomous Driving

Long-LRM: Long-sequence Large Reconstruction Model for Wide-coverage Gaussian Splats

Long Context Tuning for Video Generation

EvaGaussians: Event Stream Assisted Gaussian Splatting from Blurry Images

Lumina-Image 2.0: A Unified and Efficient Image Generative Framework

Describe Anything: Detailed Localized Image and Video Captioning

Beyond Next-Token: Next-X Prediction for Autoregressive Visual Generation

Aether: Geometric-Aware Unified World Modeling

DrivingGPT: Unifying Driving World Modeling and Planning with Multi-modal Autoregressive Transformers

MagicDrive-V2: High-Resolution Long Video Generation for Autonomous Driving with Adaptive Control

UniPortrait: A Unified Framework for Identity-Preserving Single- and Multi-Human Image Personalization

ILLUME: Illuminating Your LLMs to See, Draw, and Self-Enhance

CC-OCR: A Comprehensive and Challenging OCR Benchmark for Evaluating Large Multimodal Models in Literacy

Scaling Language-Free Visual Representation Learning

Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis

Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces

OmniGen: Unified Image Generation

CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models

Reconstruction vs. Generation: Taming Optimization Dilemma in Latent Diffusion Models

StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Text

Video-XL: Extra-Long Vision Language Model for Hour-Scale Video Understanding

GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control

ShowUI: One Vision-Language-Action Model for GUI Visual Agent

WonderWorld: Interactive 3D Scene Generation from a Single Image

From Slow Bidirectional to Fast Autoregressive Video Diffusion Models

FoundationStereo: Zero-Shot Stereo Matching

Transformers without Normalization

Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models

DEIM: DETR with Improved Matching for Fast Convergence

FLARE: Feed-forward Geometry, Appearance and Camera Estimation from Uncalibrated Sparse Views

MLVU: Benchmarking Multi-task Long Video Understanding

RoboBrain: A Unified Brain Model for Robotic Manipulation from Abstract to Concrete

DriveDreamer4D: World Models Are Effective Data Machines for 4D Driving Scene Representation

Teaching Large Language Models to Regress Accurate Image Quality Scores Using Score Distribution

MV-DUSt3R+: Single-Stage Scene Reconstruction from Sparse Views In 2 Seconds

AC3D: Analyzing and Improving 3D Camera Control in Video Diffusion Transformers

MMVU: Measuring Expert-Level Multi-Discipline Video Understanding

UniReal: Universal Image Generation and Editing via Learning Real-world Dynamics

Adaptive Keyframe Sampling for Long Video Understanding

Mono-InternVL: Pushing the Boundaries of Monolithic Multimodal Large Language Models with Endogenous Visual Pre-training

One-Minute Video Generation with Test-Time Training

SF3D: Stable Fast 3D Mesh Reconstruction with UV-unwrapping and Illumination Disentanglement

RandAR: Decoder-only Autoregressive Visual Generation in Random Orders

Go-with-the-Flow: Motion-Controllable Video Diffusion Models Using Real-Time Warped Noise

CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer

RAFT: Reward rAnked FineTuning for Generative Foundation Model Alignment

WizardMath: Empowering Mathematical Reasoning for Large Language Models via Reinforced Evol-Instruct

Causal Reasoning and Large Language Models: Opening a New Frontier for Causality

Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks

RDT-1B: a Diffusion Foundation Model for Bimanual Manipulation

Generative Verifiers: Reward Modeling as Next-Token Prediction

OpenHands: An Open Platform for AI Software Developers as Generalist Agents

Depth Pro: Sharp Monocular Metric Depth in Less Than a Second

Scaling and evaluating sparse autoencoders

Transfusion: Predict the Next Token and Diffuse Images with One Multi-Modal Model

Safety Alignment Should be Made More Than Just a Few Tokens Deep

Mixture-of-Agents Enhances Large Language Model Capabilities

Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers

MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motion

Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models

To CoT or not to CoT? Chain-of-thought helps mainly on math and symbolic reasoning

LoRA Learns Less and Forgets Less

Generating with Confidence: Uncertainty Quantification for Black-box Large Language Models

mPLUG-Owl3: Towards Long Image-Sequence Understanding in Multi-Modal Large Language Models

Generative Representational Instruction Tuning

Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge

Self-Play Preference Optimization for Language Model Alignment

Inverse Scaling: When Bigger Isn't Better

AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents

Revisiting Feature Prediction for Learning Visual Representations from Video

Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts

G-LLaVA: Solving Geometric Problem with Multi-Modal Large Language Model

The Unreasonable Ineffectiveness of the Deeper Layers

Diffusion Models Are Real-Time Game Engines