Fetching the paper…

On the Blind Spots of Model-Based Evaluation Metrics for Text Generation · Around