arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2401.03646cs.LGcs.NE
收藏

评估机制可解释性中自动电路发现的脑启发模块化训练

Evaluating Brain-Inspired Modular Training in Automated Circuit Discovery for Mechanistic Interpretability

University of Massachusetts Amherst(马萨诸塞大学阿默斯特分校)

Jatin Nainani

AI总结:

本文评估了脑启发模块化训练(BIMT)方法在自动电路发现中的应用,发现其能显著提升电路发现的效率、质量、稀疏性并克服人工局限,推动透明可信AI系统的发展。

摘要:

大型语言模型(LLMs)在人工智能领域迅速崛起,以其先进的能力改变了广泛的应用。随着这些模型日益成为决策的核心,对彻底的可解释性的需求变得前所未有的关键。机制可解释性通过识别和分析这些复杂系统中的特定子网络或“电路”,为这种理解提供了一条途径。这种方法的一个关键方面是自动电路发现,它使得研究像 GPT4 或 LLAMA 这样的大型模型变得可行。在此背景下,我们的研究评估了一种旨在增强神经网络可解释性的最新方法——脑启发模块化训练(BIMT)。我们展示了 BIMT 如何显著提升自动电路发现的效率和质量,克服了人工方法的局限性。我们的比较分析进一步揭示,BIMT 在电路质量、发现时间和稀疏性方面优于现有模型。此外,我们还对 BIMT 进行了全面的计算分析,包括训练时长、内存分配需求和推理速度等方面。除了展示 BIMT 在使神经网络更易于理解方面的良好效果外,本研究还推动了创建可信赖和透明的人工智能系统的更大目标。

英文摘要:

Large Language Models (LLMs) have experienced a rapid rise in AI, changing a wide range of applications with their advanced capabilities. As these models become increasingly integral to decision-making, the need for thorough interpretability has never been more critical. Mechanistic Interpretability offers a pathway to this understanding by identifying and analyzing specific sub-networks or 'circuits' within these complex systems. A crucial aspect of this approach is Automated Circuit Discovery, which facilitates the study of large models like GPT4 or LLAMA in a feasible manner. In this context, our research evaluates a recent method, Brain-Inspired Modular Training (BIMT), designed to enhance the interpretability of neural networks. We demonstrate how BIMT significantly improves the efficiency and quality of Automated Circuit Discovery, overcoming the limitations of manual methods. Our comparative analysis further reveals that BIMT outperforms existing models in terms of circuit quality, discovery time, and sparsity. Additionally, we provide a comprehensive computational analysis of BIMT, including aspects such as training duration, memory allocation requirements, and inference speed. This study advances the larger objective of creating trustworthy and transparent AI systems in addition to demonstrating how well BIMT works to make neural networks easier to understand.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑