Limitations of Language Models in Arithmetic and Symbolic Induction

The dataset used in the paper to test the limitations of large pretrained Language Models (LMs) on arithmetic and symbolic induction tasks.

BibTex: