Dan Zhang

TDRM: Smooth Reward Models with Temporal Difference for LLM RL and Inference

ReST-RL: Reinforcing LLM Reasoning through Self-Training and Value-Guided Decoding

A Stronger Mixture of Low-Rank Experts for Fine-Tuning Foundation Models

SPaR: Self-Play with Tree-Search Refinement to Improve Instruction-Following in Large Language Models

Can Large Language Models Master Complex Card Games?

ZeroFlow: Overcoming Catastrophic Forgetting is Easier than You Think

Self-Control of LLM Behaviors by Compressing Suffix Gradient into Prefix Controller

DataSciBench: An LLM Agent Benchmark for Data Science