SimpleLART is a quickly created dataset inspired by the paper How well do SOTA legal reasoning models support abductive reasoning?. It is designed to simplify the process of working with datasets without requiring complex libraries or multilingual setups. SimpleLART shares characteristics and the number of samples similar to the L'ART dataset described in the paper, with the following notes:
Instead of generating samples randomly via theory generation using Lisp… See the full description on the dataset page:
https://huggingface.co/datasets/nguyenthanhasia/SimpleLART.
