@inproceedings{chang-etal-2025-mind,
    title = "Mind the Gap: How {B}aby{LM}s Learn Filler-Gap Dependencies",
    author = "Chang, Chi-Yun  and
      Huang, Xueyang  and
      Nasir, Humaira  and
      Storks, Shane  and
      Akingbade, Olawale  and
      Dai, Huteng",
    editor = "Christodoulopoulos, Christos  and
      Chakraborty, Tanmoy  and
      Rose, Carolyn  and
      Peng, Violet",
    booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2025.emnlp-main.761/",
    pages = "15060--15076",
    ISBN = "979-8-89176-332-6",
    abstract = "Humans acquire syntactic constructions like filler-gap dependencies from limited and often noisy input. Can neural language models do the same? We investigate this question by evaluating GPT-2 models trained on child-oriented input from the BabyLM Challenge. Our experiments focus on whether these ``baby'' language models acquire filler-gap dependencies, generalize across constructions, and respect structural constraints such as island effects. We apply a suite of syntactic constructions to four models trained on child language, including two base models (trained on 10M and 100M tokens) and two well-performing models from the BabyLM Challenge (ConcreteGPT and BabbleGPT). We evaluate model behavior using wh-licensing scores, flip tests, and grammaticality contrasts across four constructions. Results show that BabyLM-scale models partially acquire filler-gap dependencies but often fail to generalize or fully capture island constraints."
}