← authors · overview

• Defeating the Training-Inference Mismatch via FP16
  2025 · 34 citations
• LeTS-Drive: Driving in a Crowd by Learning from Tree Search
  2019 · 34 citations
• Rethinking the Divergence Regularization in LLM RL
  2026
• EUBRL: Epistemic Uncertainty Directed Bayesian Reinforcement Learning
  2025
• SofT-GRPO: Surpassing Discrete-Token LLM Reinforcement Learning via Gumbel-Reparameterized Soft-Thinking Policy Optimization
  2025
• GEM: A Gym for Agentic LLMs
  2025
• SofT-GRPO: Surpassing Discrete-Token LLM Reinforcement Learning via Gumbel-Reparameterized Soft-Thinking Policy Optimization
  2025
• Is Model Editing Built on Sand? Revealing Its Illusory Success and Fragile Foundation
  2025
• Defeating the Training-Inference Mismatch via FP16
  2025
• Extending Epistemic Uncertainty Beyond Parameters Would Assist in Designing Reliable LLMs
  2025
• SPIRAL: Self-Play on Zero-Sum Games Incentivizes Reasoning via Multi-Agent Multi-Turn Reinforcement Learning
  2025
• First Try Matters: Revisiting the Role of Reflection in Reasoning Models
  2025
• SofT-GRPO: Surpassing Discrete-Token LLM Reinforcement Learning via Gumbel-Reparameterized Soft-Thinking Policy Optimization
  2025
• Sample-Efficient Alignment for LLMs
  2024