BabyBabelLM: A Multilingual Benchmark of Developmentally Plausible Training Data
Jumelet, Jaap, Fourtassi, Abdellah, Haga, Akari, Bunzeck, Bastian, Shandilya, Bhargav, Galvan-Sosa, Diana, Haznitrama, Faiz Ghifari, Padovani, Francesca, Meyer, Francois, Hu, Hai, Etxaniz, Julen, Prévot, Laurent, He, Linyang, Grandury, María, Marcheva, Mila, Foroutan, Negar, Theodoropoulos, Nikitas, Sadeghi, Pouya, Song, Siyuan, Salhan, Suchir, Zhou, Susana, Paniv, Yurii, Zhang, Ziyin, Bisazza, Arianna, Warstadt, Alex, Choshen, Leshem
–arXiv.org Artificial Intelligence
We present BabyBabelLM, a multilingual collection of datasets modeling the language a person observes from birth until they acquire a native language. We curate developmentally plausible pretraining data aiming to cover the equivalent of 100M English words of content in each of 45 languages. We compile evaluation suites and train baseline models in each language. BabyBabelLM aims to facilitate multilingual pretraining and cognitive modeling.
arXiv.org Artificial Intelligence
Oct-14-2025
- Country:
- North America > United States (1.00)
- Europe (1.00)
- Asia (1.00)
- Africa (0.67)
- Genre:
- Research Report > New Finding (0.67)
- Industry:
- Media (1.00)
- Government (0.93)
- Leisure & Entertainment (0.67)
- Education > Educational Setting
- K-12 Education (0.68)
- Technology:
- Information Technology > Artificial Intelligence
- Representation & Reasoning (1.00)
- Machine Learning (1.00)
- Natural Language > Large Language Model (0.47)
- Speech > Speech Recognition (0.45)
- Information Technology > Artificial Intelligence