Reproduce “Dango: A Strictly L1-Only Large Language Model for Studying Second Language Acquisition” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Shiho Matta, Yin Jou Huang, Fei Cheng, Takashi Kodama, Hirokazu Kiyomaru, Yugo Murawaki
Dango trains a 1.8B-parameter decoder model first on filtered Japanese web text, retaining short English phrases, then introduces English through Japanese-language lessons. The authors compare generated English with Japanese learners’ writing across seven feature categories. Dango’s error-rate JSD is 0.0110 versus 0.0304 for an unfiltered baseline, but it is not closest on every metric or proof of human-like acquisition.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.