Pengyu Cheng

Eliminating Inductive Bias in Reward Models with Information-Theoretic Guidance

Search Self-Play: Pushing the Frontier of Agent Capability without Supervision

Self-playing Adversarial Language Game Enhances LLM Reasoning

Chunk, Align, Select: A Simple Long-sequence Processing Method for Transformers

Everyone Deserves A Reward: Learning Customized Human Preferences