摘要：理解大型语言模型（LLMs）中的语言习得动态，对于揭示统计表征如何演化为结构化语言能力至关重要。然而，现有大多数研究主要关注已完成训练的模型，因而对中间训练阶段的发展轨迹探索不足。在本研究中，我们利用开源模型（如 Pythia 和 OLMo）的中间检查点，在 teacher-forcing 评估设置下，系统研究不同训练步数和模型规模下的语言习得。我们聚焦四项核心特征：形态结构习得、嵌入漂移、重复主导效应和共现效应。研究发现，模型在语义嵌入稳定之前很早便习得了形态结构。此外，我们表明，训练数据集规模会对分歧识别产生积极影响；而预训练期间的共现，以及推理期间的重复，则会对分歧识别和关联链接生成产生负面影响。

Abstract:Understanding the dynamics of language acquisition in Large Language Models (LLMs) is crucial for uncovering how statistical representations evolve into structured linguistic capabilities. However, most existing studies focus primarily on fully trained models, leaving the developmental trajectory across intermediate training stages largely underexplored. In this work, we leverage intermediate checkpoints from open-source models (e.g., Pythia and OLMo) under teacher-forcing evaluation to systematically investigate language acquisition across training steps and model scales. We focus on four core characteristics: morphological structure acquisition, embedding drift, repetition dominance, and co-occurrence effects. Our findings reveal that models acquire morphological structures well before semantic embeddings stabilize. Furthermore, we show that training dataset size has a positive impact on divergence recognition, while co-occurrence during pre-training and repetition during inference have a negative impact on divergence recognition and associative link generation.