A dataset for Natural Language Inference in Arabic. This dataset was compiled in the scope of the Bachelor thesis by Majd Saad al Deen (HS Bonn Rhein-Sieg) with the title "Informierte Pre-Training Methoden für Natural Language Inference im Arabischen". It consists of samples from the SNLI, XNLI and arNLI data sets. Further information can be found in the paper "Improving Natural Language Inference in Arabic using Transformer Models and Linguistically Informed Pre-Training" (submitted to IEEE SSCI 2023).