Reinforcement Learning with Token-level Feedback for Controllable Text Generation.

Wendi Li, Wei Wei, Kaihe Xu, Wenfeng Xie, Dangyang Chen, Yu Cheng

Browse the full NAACL paper archive.