用复杂网络分析大模型训练中的功能电路形成过程
Modeling Transformers as complex networks to analyze learning dynamics
- 将Transformer组件建模为有向加权图，通过干预消融量化因果影响
- 发现训练分探索、巩固、精炼三阶段，信息传播者角色稳定而收集者动态变化
- 适合关注大模型内部机制演化与可解释性研究的读者
大型语言模型（LLM）在训练过程中如何获得复杂能力，仍是机械可解释性领域的关键开放问题。本研究探讨是否可通过复杂网络理论（CNT）来刻画这一学习动态。提出一种新方法，将基于Transformer的LLM表示为有向加权图，节点为计算组件（注意力头和MLP），边表示因果影响，采用基于干预的消融技术进行度量。通过追踪Pythia-14M模型在标准归纳任务上143个训练检查点的组件图演变，分析一系列图论指标。结果表明，网络结构经历显著的探索、巩固与精炼阶段。具体而言，识别出一个稳定的‘信息传播者’层级结构，以及一组在关键学习节点动态重组的‘信息收集者’组件。该研究证明，组件级网络视角为可视化和理解LLM中功能性电路形成的自组织原则提供了有力的宏观视角。
原文摘要 · Abstract (English)
The process by which Large Language Models (LLMs) acquire complex capabilities during training remains a key open question in mechanistic interpretability. This project investigates whether these learning dynamics can be characterized through the lens of Complex Network Theory (CNT). I introduce a novel methodology to represent a Transformer-based LLM as a directed, weighted graph where nodes are the model's computational components (attention heads and MLPs) and edges represent causal influence, measured via an intervention-based ablation technique. By tracking the evolution of this component-graph across 143 training checkpoints of the Pythia-14M model on a canonical induction task, I analyze a suite of graph-theoretic metrics. The results reveal that the network's structure evolves through distinct phases of exploration, consolidation, and refinement. Specifically, I identify the emergence of a stable hierarchy of information spreader components and a dynamic set of information gatherer components, whose roles reconfigure at key learning junctures. This work demonstrates that a component-level network perspective offers a powerful macroscopic lens for visualizing and understanding the self-organizing principles that drive the formation of functional circuits in LLMs.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。