Data Annotation, Model Debugging, and Data Work / NLP Model Error Analysis and Visualization Support

How can multi-level semantic features (e.g., lexical, conceptual, high-level features) explain NLP model error distributions and behavior?

Similar questions

Related papers