Lil-Bevo: Explorations of Strategies for Training Language Models in More Humanlike Ways