Xuekai Zhu

SimpleVLA-RL: Scaling VLA Training via Reinforcement Learning

FlowRL: Matching Reward Distributions for LLM Reasoning

Flow of Spans: Generalizing Language Models to Dynamic Span-Vocabulary via GFlowNets

Towards a Unified View of Large Language Model Post-Training

Reasoning via Test-Time Instance-Level Policy Gradient in Latent Space

How Far Can Unsupervised RLVR Scale LLM Training?

MARTI: A Framework for Multi-Agent LLM Systems Reinforced Training and Inference

SEMA: Simple yet Effective Learning for Multi-Turn Jailbreak Attacks

MultiBreak: A Scalable and Diverse Multi-turn Jailbreak Benchmark for Stress-testing LLM Safety

SSRL: Self-Search Reinforcement Learning

$\alpha$-GFN: Generalized Mixing in GFlowNets for Better Exploration-Exploitation Trade-off

TTRL: Test-Time Reinforcement Learning

MedXpertQA: Benchmarking Expert-Level Medical Reasoning and Understanding

ToEdit: How to Synthesize Text Data to Avoid Model Collapse?

OpenPRM: Building Open-domain Process-based Reward Models with Preference Trees

How to Synthesize Text Data without Model Collapse?

Fourier Position Embedding: Enhancing Attention’s Periodic Extension for Length Generalization