The language model trained on a fill-mask task with all the North American parent's data in CHILDES. The parent's data can be found here: https://github.com/xiaomeng-ma/CHILDES