@inproceedings{hakimi-etal-2025-time,
    title = "Time Course {M}ech{I}nterp: Analyzing the Evolution of Components and Knowledge in Large Language Models",
    author = "Hakimi, Ahmad Dawar  and
      Modarressi, Ali  and
      Wicke, Philipp  and
      Schuetze, Hinrich",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/corrections-2025-08/2025.findings-acl.654/",
    doi = "10.18653/v1/2025.findings-acl.654",
    pages = "12633--12653",
    ISBN = "979-8-89176-256-5",
    abstract = "Understanding how large language models (LLMs) acquire and store factual knowledge is crucial for enhancing their interpretability, reliability, and efficiency. In this work, we analyze the evolution of factual knowledge representation in the OLMo-7B model by tracking the roles of its Attention Heads and Feed Forward Networks (FFNs) over training. We classify these components into four roles{---}general, entity, relation-answer, and fact-answer specific{---}and examine their stability and transitions. Our results show that LLMs initially depend on broad, general-purpose components, which later specialize as training progresses. Once the model reliably predicts answers, some components are repurposed, suggesting an adaptive learning process. Notably, answer-specific attention heads display the highest turnover, whereas FFNs remain stable, continually refining stored knowledge. These insights offer a mechanistic view of knowledge formation in LLMs and have implications for model pruning, optimization, and transparency."
}