A GPT-2 model for the BabyLM 2026 Multilingual track. It is trained with our
code-switched curriculum corpus.
This is our leaderboard submission with full checkpoints. Please see our full set of models for the paper
here
The paper, repository, and further details are TBA on release.