Abstract: This paper introduces a method for trajectory selection using large-scale pre-trained language models, aiming to improve sample and training efficiency in reinforcement learning. By using a ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback