arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2304.09707cs.CVcs.LG
收藏

使用概念向量解耦神经元表征

Disentangling Neuron Representations with Concept Vectors

Haute école spécialisée de Suisse occidentale(瑞士西部应用科学大学);University of Limerick(利默里克大学);The Sense Research and Innovation Center(The Sense研究与创新中心)

Laura O'Mahony, Vincent Andrearczyk, Henning Muller, Mara Graziani

AI总结:

针对多语义神经元难以解释的问题,本文提出一种将其解耦为不同概念向量的方法,支持按用户指定粒度搜索,并验证所得方向具有连贯、可理解的特征。

摘要:

机制可解释性旨在通过将神经网络分解为可解释单元,来理解模型如何存储表征。然而,多语义神经元(即对多个互不相关特征作出响应的神经元)的出现,使解释单个神经元变得困难。这促使研究者在激活空间而非单个神经元中寻找被称为概念向量的有意义向量。本文的主要贡献是提出一种将多语义神经元解耦为封装不同特征的概念向量的方法。该方法可根据用户期望的概念分离程度搜索细粒度概念。分析表明,多语义神经元可以被解耦为由神经元线性组合构成的方向。评估结果显示,所发现的概念向量编码了连贯且人类可理解的特征。

英文摘要:

Mechanistic interpretability aims to understand how models store representations by breaking down neural networks into interpretable units. However, the occurrence of polysemantic neurons, or neurons that respond to multiple unrelated features, makes interpreting individual neurons challenging. This has led to the search for meaningful vectors, known as concept vectors, in activation space instead of individual neurons. The main contribution of this paper is a method to disentangle polysemantic neurons into concept vectors encapsulating distinct features. Our method can search for fine-grained concepts according to the user's desired level of concept separation. The analysis shows that polysemantic neurons can be disentangled into directions consisting of linear combinations of neurons. Our evaluations show that the concept vectors found encode coherent, human-understandable features.

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。

↑