A non-saturating, open-ended environment for evaluating LLMs in Factorio
Unsupervised Language Modeling at scale for robust sentiment classification