Enhancing Post-Hoc Attributions in Long Document Comprehension via Coarse Grained Answer Decomposition

Open in new window