Recomposer: Event-roll-guided generative audio editing

Transition Models: Rethinking the Generative Learning Objective

Inverse IFEval: Can LLMs Unlearn Stubborn Training Conventions to Follow Real Instructions?

DeepResearch Arena: The First Exam of LLMs' Research Abilities via Seminar-Grounded Tasks

Towards a Unified View of Large Language Model Post-Training

From Editor to Dense Geometry Estimator

Drivel-ology: Challenging LLMs with Interpreting Nonsense with Depth

Loong: Synthesize Long Chain-of-Thoughts at Scale through Verifiers

ArcMemo: Abstract Reasoning Composition with Lifelong LLM Memory

CoT-Space: A Theoretical Framework for Internal Slow-Thinking via Reinforcement Learning

Multi-View 3D Point Tracking

MOSAIC: Multi-Subject Personalized Generation via Correspondence-Aware Alignment and Disentanglement

Mixture of Global and Local Experts with Diffusion Transformer for Controllable Face Generation

On the Theoretical Limitations of Embedding-Based Retrieval

LMEnt: A Suite for Analyzing Knowledge in Language Models from Pretraining Data to Representations

Open Data Synthesis For Deep Research

Robix: A Unified Model for Robot Interaction, Reasoning and Planning

FusionProt: Fusing Sequence and Structural Information for Unified Protein Representation Learning

LimiX: Unleashing Structured-Data Modeling Capability for Generalist Intelligence

epiGPTope: A machine learning-based epitope generator and classifier

GenCompositor: Generative Video Compositing with Diffusion Transformer

DCPO: Dynamic Clipping Policy Optimization

Reasoning Vectors: Transferring Chain-of-Thought Capabilities via Task Arithmetic

Baichuan-M2: Scaling Medical Capability with Large Verifier System

VerlTool: Towards Holistic Agentic Reinforcement Learning with Tool Use

ELV-Halluc: Benchmarking Semantic Aggregation Hallucinations in Long Video Understanding

AlphaEarth Foundations: An embedding field model for accurate and efficient global mapping from sparse label data

AetherCode: Evaluating LLMs' Ability to Win In Premier Programming Competitions

TileLang: A Composable Tiled Programming Model for AI Systems

DeepSeek-R1 Thoughtology: Let's think about LLM Reasoning

Recomposer: Event-roll-guided generative audio editing

Transition Models: Rethinking the Generative Learning Objective

Inverse IFEval: Can LLMs Unlearn Stubborn Training Conventions to Follow Real Instructions?

DeepResearch Arena: The First Exam of LLMs' Research Abilities via Seminar-Grounded Tasks

Towards a Unified View of Large Language Model Post-Training

From Editor to Dense Geometry Estimator

Drivel-ology: Challenging LLMs with Interpreting Nonsense with Depth

Loong: Synthesize Long Chain-of-Thoughts at Scale through Verifiers

ArcMemo: Abstract Reasoning Composition with Lifelong LLM Memory

CoT-Space: A Theoretical Framework for Internal Slow-Thinking via Reinforcement Learning

Multi-View 3D Point Tracking

MOSAIC: Multi-Subject Personalized Generation via Correspondence-Aware Alignment and Disentanglement

Mixture of Global and Local Experts with Diffusion Transformer for Controllable Face Generation

On the Theoretical Limitations of Embedding-Based Retrieval

LMEnt: A Suite for Analyzing Knowledge in Language Models from Pretraining Data to Representations

Open Data Synthesis For Deep Research

Robix: A Unified Model for Robot Interaction, Reasoning and Planning

FusionProt: Fusing Sequence and Structural Information for Unified Protein Representation Learning

LimiX: Unleashing Structured-Data Modeling Capability for Generalist Intelligence

epiGPTope: A machine learning-based epitope generator and classifier

GenCompositor: Generative Video Compositing with Diffusion Transformer

DCPO: Dynamic Clipping Policy Optimization

Reasoning Vectors: Transferring Chain-of-Thought Capabilities via Task Arithmetic

Baichuan-M2: Scaling Medical Capability with Large Verifier System

VerlTool: Towards Holistic Agentic Reinforcement Learning with Tool Use

ELV-Halluc: Benchmarking Semantic Aggregation Hallucinations in Long Video Understanding

AlphaEarth Foundations: An embedding field model for accurate and efficient global mapping from sparse label data

AetherCode: Evaluating LLMs' Ability to Win In Premier Programming Competitions

TileLang: A Composable Tiled Programming Model for AI Systems

DeepSeek-R1 Thoughtology: Let's think about LLM Reasoning