1 paper
Tianyi Zhang, Mina Lee, Lisa Li +2
While pretrained language models (PLMs) have greatly improved text generation, they have also been known to produce unfaithful or inappropriate content. In contrast, classic templa…