Models for text generation have become focal for many research tasks and especially for the generation of sentence corpora. However, understanding the properties of an automatically generated text corpus remains challenging. We propose a set of tools that examine the properties of generated text corpora. Applying these tools on various generated corpora allowed us to gain new insights into the properties of the generative models. As part of our characterization process, we found remarkable differences in the corpora generated by two leading generative technologies.
@article{arxiv.2206.11219,
title = {Understanding the Properties of Generated Corpora},
author = {Naama Zwerdling and Segev Shlomov and Esther Goldbraich and George Kour and Boaz Carmeli and Naama Tepper and Inbal Ronen and Vitaly Zabershinsky and Ateret Anaby-Tavor},
journal= {arXiv preprint arXiv:2206.11219},
year = {2022}
}