Zeming Li
2 indexed papers
Publications per year
Top categories
Frequent co-authors
Research Timeline
ESPO is a novel reinforcement learning algorithm that detects trajectory failure in large language models and terminates rollouts early, significantly improving performance on mathematical reasoning benchmarks while reducing computational cost.
ESPO is a novel reinforcement learning algorithm that detects trajectory failure in large language models and terminates rollouts early, significantly improving performance on mathematical reasoning benchmarks while reducing computational cost.
Papers
ESPO: Early-Stopping Proximal Policy Optimization
Zihang Li, Rui Zhou, Yingcheng Shi, Wenhan Yu +7 more
ESPO is a novel reinforcement learning algorithm that detects trajectory failure in large language models and terminates rollouts early, significantly improving performance on mathematical reasoning b…