Guanhua Huang

ArtifactsBench: Bridging the Visual-Interactive Gap in LLM Code Generation Evaluation

Segmental Advantage Estimation: Enhancing PPO for Long-Context LLM Training

Low-probability Tokens Sustain Exploration in Reinforcement Learning with Verifiable Reward

Improving Generalization and Convergence by Enhancing Implicit Regularization

AGILE: A Novel Reinforcement Learning Framework of LLM Agents