@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://aclanthology.org/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."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="hakimi-etal-2025-time">
<titleInfo>
<title>Time Course MechInterp: Analyzing the Evolution of Components and Knowledge in Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ahmad</namePart>
<namePart type="given">Dawar</namePart>
<namePart type="family">Hakimi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ali</namePart>
<namePart type="family">Modarressi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Philipp</namePart>
<namePart type="family">Wicke</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hinrich</namePart>
<namePart type="family">Schuetze</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-256-5</identifier>
</relatedItem>
<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.</abstract>
<identifier type="citekey">hakimi-etal-2025-time</identifier>
<identifier type="doi">10.18653/v1/2025.findings-acl.654</identifier>
<location>
<url>https://aclanthology.org/2025.findings-acl.654/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>12633</start>
<end>12653</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Time Course MechInterp: Analyzing the Evolution of Components and Knowledge in Large Language Models
%A Hakimi, Ahmad Dawar
%A Modarressi, Ali
%A Wicke, Philipp
%A Schuetze, Hinrich
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F hakimi-etal-2025-time
%X 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.
%R 10.18653/v1/2025.findings-acl.654
%U https://aclanthology.org/2025.findings-acl.654/
%U https://doi.org/10.18653/v1/2025.findings-acl.654
%P 12633-12653
Markdown (Informal)
[Time Course MechInterp: Analyzing the Evolution of Components and Knowledge in Large Language Models](https://aclanthology.org/2025.findings-acl.654/) (Hakimi et al., Findings 2025)
ACL