Fetching the paper…

Reinforcement Learning with Token-level Feedback for Controllable Text Generation · Around