Shao Tang

Reasoning Models Can be Accurately Pruned Via Chain-of-Thought Reconstruction

Planner-R1: Reward Shaping Enables Efficient Agentic RL with Smaller LLMs

On quantizing the state of the Muon optimizer

In-Context Bayesian Reward Modeling for Test-Time Steerability

LLM Query Scheduling with Prefix Reuse and Latency Constraints

AlphaPO: Reward Shape Matters for LLM Alignment

On the Robustness of Reward Models for Language Model Alignment