Interpreting Pretrained Language Models via Concept Bottlenecks
Abstract
Pretrained language models (PLMs) have made significant strides in various natural language processing tasks. However, the lack of interpretability due to their ``black-box'' nature poses challenges for responsible implementation. Although previous studies have attempted to improve interpretability by using, e.g., attention weights in self-attention layers, these weights often lack clarity, readability, and intuitiveness. In this research, we propose a novel approach to interpreting PLMs by employing high-level, meaningful concepts that are easily understandable for humans. For example, we learn the concept of ``Food'' and investigate how it influences the prediction of a model's sentiment towards a restaurant review. We introduce CM, which combines human-annotated and machine-generated concepts to extract hidden neurons designed to encapsulate semantically meaningful and task-specific concepts. Through empirical evaluations on real-world datasets, we manifest that our approach offers valuable insights to interpret PLM behavior, helps diagnose model failures, and enhances model robustness amidst noisy concept labels.
Cite
@article{arxiv.2311.05014,
title = {Interpreting Pretrained Language Models via Concept Bottlenecks},
author = {Zhen Tan and Lu Cheng and Song Wang and Yuan Bo and Jundong Li and Huan Liu},
journal= {arXiv preprint arXiv:2311.05014},
year = {2023}
}