LittleLearner: Language Models Under Pedagogically Controlled Knowledge Exposure

LITTLECURRICULUM contains **88B tokens** aligned to U.S. elementary material while excluding concepts, facts, and vocabulary taught above Grade 5. A **5B-parameter** model trained from scratch provides enough language ability for open-ended evaluation within those boundaries.

This gives researchers a cleaner sandbox for testing post-training, prompting, and in-context learning. Builders evaluating adaptation methods can use the setup to distinguish retrieval or recombination of known material from acquisition beyond the training scope.

The first experiments found better use of existing knowledge but **no increase in out-of-scope capabilities**. The material does not establish whether that result generalizes to broader corpora, larger models, or agent workflows.