ResearcharXivNEW

LittleLearner: Language Models Under Pedagogically Controlled Knowledge Exposure

Li 2026-08-13

Fanfei LiJana ZellerManuel Prada-Corral

Modern language models are trained on heterogeneous web-scale text corpora. Consequently, studying knowledge and skill acquisition is difficult, as prior exposure to related content is hard to characterize. To address this challenge, we introduce LITTLECURRICULUM, a curated 88B-token pretraining corpus tailored to U.S. elementary school material, explicitly excluding concepts, facts, and vocabular

Read on arXiv

Data aggregated and editorially reviewed by TrendMing.

Key Contributions

Modern language models are trained on heterogeneous web-scale text corpora.
Consequently, studying knowledge and skill acquisition is difficult, as prior exposure to related content is hard to characterize.
To address this challenge, we introduce LITTLECURRICULUM, a curated 88B-token pretraining corpus tailored to U.S.

Research Themes

AIResearch

Back to AI Research