Ge Zhang

In-Place Test-Time Training

FutureX: An Advanced Live Benchmark for LLM Agents in Future Prediction

Reformulation for Pretraining Data Augmentation

YuE: Scaling Open Foundation Models for Long-Form Music Generation

WideSearch: Benchmarking Agentic Broad Info-Seeking

Reverse-Engineered Reasoning for Open-Ended Generation

ScaleLong: A Multi-Timescale Benchmark for Long Video Understanding

IV-Bench: A Benchmark for Image-Grounded Video Perception and Reasoning in Multimodal LLMs

FinSearchComp: Towards a Realistic, Expert-Level Evaluation of Financial Search and Reasoning

Flash-Searcher: Fast and Effective Web Agents via DAG-Based Parallel Execution

TaskCraft: Automated Generation of Agentic Tasks

TreePO: Enhancing Policy Efficacy and Inference Efficiency with Tree Modeling

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

FormalMATH: Benchmarking Formal Mathematical Reasoning of Large Language Models

ReTool: Reinforcement Learning for Strategic Tool Use in LLMs

P2P: Automated Paper-to-Poster Generation and Fine-Grained Benchmark

A$^2$FM: An Adaptive Agent Foundation Model for Tool-Aware Hybrid Reasoning

VeriWeb: Verifiable Long-Chain Web Benchmark for Agentic Information-Seeking

DiscoX: Benchmarking Discourse-Level Translation in Expert Domains

Chain-of-Agents: End-to-End Agent Foundation Models via Multi-Agent Distillation and Agentic RL

Knapsack RL: Unlocking Exploration of LLMs via Optimizing Budget Allocation

ACADREASON: Exploring the Limits of Reasoning Models with Academic Research Problems

OmniVideoBench: Towards Audio-Visual Understanding Evaluation for Omni MLLMs

AttentionInfluence: Adopting Attention Head Influence for Weak-to-Strong Pretraining Data Selection

Beyond Score: A Multi-Agent System to Discover Capability and Behavioral Weaknesses in LLMs

Towards Personalized Deep Research: Benchmarks and Evaluations

MM-BrowseComp: A Comprehensive Benchmark for Multimodal Browsing Agents

Beyond Correctness: Evaluating Subjective Writing Preferences Across Cultures

SciDA: Scientific Dynamic Assessor of LLMs

Audio-FLAN: An Instruction-Following Dataset for Unified Understanding and Generation of Speech, Music, and Sound

VideoEval-Pro: Robust and Realistic Long Video Understanding Evaluation

CryptoX : Compositional Reasoning Evaluation of Large Language Models

First return, entropy-eliciting explore

MME-CC: A Challenging Multi-Modal Evaluation Benchmark of Cognitive Capacity

VideoScore2: Think before You Score in Generative Video Evaluation

COIG-Writer: A High-Quality Chinese Creative Writing with Thought Process Dataset

LPFQA: A Long-Tail Professional Forum-based Benchmark for LLMs' Evaluation

KORGym: A Dynamic Game Platform for LLM Reasoning Evaluation

KOR-Bench: Benchmarking Language Models on Knowledge-Orthogonal Reasoning Tasks

FlexWorld: Progressively Expanding 3D Scenes for Flexible-View Exploration

Omni-MATH: A Universal Olympiad Level Mathematic Benchmark for Large Language Models

VCR: A Task for Pixel-Level Complex Reasoning in Vision Language Models via Restoring Occluded Text

McEval: Massively Multilingual Code Evaluation

MuPT: A Generative Symbolic Music Pretrained Transformer

General-Reasoner: Advancing LLM Reasoning Across All Domains

LIME: LESS IS MORE FOR MLLM EVALUATION

OmniEdit: Building Image Editing Generalist Models Through Specialist Supervision

MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark

OmniBench: Towards The Future of Universal Omni-Language Models

MTU-Bench: A Multi-granularity Tool-Use Benchmark for Large Language Models

M2rc-Eval: Massively Multilingual Repository-level Code Completion Evaluation

MIO: A Foundation Model on Multimodal Tokens

AutoKaggle: A Multi-Agent Framework for Autonomous Data Science Competitions

HelloBench: Evaluating Long Text Generation Capabilities of Large Language Models

KARPA: A Training-free Method of Adapting Knowledge Graph as References for Large Language Model's Reasoning Path Aggregation

ING-VP: MLLMs Cannot Play Easy Vision-based Games Yet

Can MLLMs Understand the Deep Implication Behind Chinese Images?