The XNLI dataset comprises pairs of sentences with a label categorizing the semantic relationship between the two sentences into one of three classifications: entailment,...
The dataset used in the paper is the Stanford Natural Language Inference (SNLI) dataset, which consists of 549,367 premise-hypothesis pairs for train/dev/test sets and target...
