Wei Xiong

StepWiser: Stepwise Generative Judges for Wiser Reasoning

Logarithmic Regret for Online KL-Regularized Reinforcement Learning

Building Math Agents with Multi-Turn Iterative Preference Learning

Optimizing Chain-of-Thought Reasoners via Gradient Variance Minimization in Rejection Sampling and RL

LLM Alignment as Retriever Optimization: An Information Retrieval Perspective

RRM: Robust Reward Model Training Mitigates Reward Hacking

DPO Meets PPO: Reinforced Token Optimization for RLHF