paper-with-me

FED

Fine-grained Evaluation of Dialog

· 논문 53편

The FED dataset is constructed by annotating a set of human-system and human-human conversations with eighteen fine-grained dialog qualities.