SyntaxShap datasets

This project contains the three datasets used in the paper <i>"Syntax-aware Explainability Method for Text Generation".</i>-<b><i> </i></b><b><i>Inconsistent Negation Dataset</i></b>: contains disjoint sentence pairs, i.e., a sentence and its negated version, whereby the last token in the sentence is masked with the [MASK] token;- <b><i>Generics KB</i></b>: contains generic sentences, which are high-quality, semantically complete statements;- <b><i>ROCStories Winter2017</i></b>: contains a collection of five-sentence everyday life stories.<br>More information on the preprocessing and filtering of the data can be found in the paper.