2 datasets found

SlimPajama: A 627B token cleaned and deduplicated version of RedPajama.

The dataset is used to evaluate the performance of the xLSTM architecture on various tasks, including language modeling, question answering, and text classification.