Exploring Text Classification Models with Sparse Autoencoders

2026-09-17

AI Summary

本文使用稀疏自动编码器分析文本分类模型的行为，通过探索特征与模型预测及错误间的关系来提高模型透明度，并开发了SAEfarer工具。

Abstract

As language models (LMs) rise in prominence, there is interest in making them more transparent in order to better understand their internal behavior. Recent interpretability work has focused on using sparse autoencoders (SAEs) to break down neuron activations at a given layer in the LM into human-understandable features, where each feature represents a concept that the model has learned. In this paper, we share work on using SAEs to analyze the behavior of text classification LMs. We present techniques for exploring the relationships between the SAE's features and the model's predictions and errors. We integrate these techniques into SAEfarer, a tool for analyzing concepts learned by text classification LMs. We assess SAEfarer in an expert pilot evaluation with five Ph.D. students.

Problem

• language models
• transparency
• internal behavior
• text classification
• sparse autoencoders

Innovation

• Sparse Autoencoders
• Text Classification Models
• Interpretability
• SAEfarer