Zejun MA

SimpleTIR: End-to-End Reinforcement Learning for Multi-Turn Tool-Integrated Reasoning

video-SALMONN S: Streaming Audio-Visual LLMs Beyond Length Limits via Memory

video-SALMONN 2: Caption-Enhanced Audio-Visual Large Language Models

SWE-Perf: Can Language Models Optimize Code Performance on Real-World Repositories?

First return, entropy-eliciting explore

Robust SuperAlignment: Weak-to-Strong Robustness Generalization for Vision-Language Models

LLaVA-Interleave: Tackling Multi-image, Video, and 3D in Large Multimodal Models

Afterburner: Reinforcement Learning Facilitates Self-Improving Code Efficiency Optimization

SimpleRL-Zoo: Investigating and Taming Zero Reinforcement Learning for Open Base Models in the Wild

VisionThink: Smart and Efficient Vision Language Model via Reinforcement Learning

ZeCO: Zero-Communication Overhead Sequence Parallelism for Linear Attention

General-Reasoner: Advancing LLM Reasoning Across All Domains

Improving LLM Video Understanding with 16 Frames Per Second

video-SALMONN-o1: Reasoning-enhanced Audio-visual Large Language Model

Video Instruction Tuning with Synthetic Data

Enhancing Multimodal LLM for Detailed and Accurate Video Captioning using Multi-Round Preference Optimization