Investigating the Robustness of Natural Language Generation from Logical Forms via Counterfactual Samples

Open in new window