Scaling up Discovery of Latent Concepts in Deep NLP Models

TL;DR

This paper improves the scalability of discovering latent concepts in deep NLP models by evaluating clustering algorithms, especially K-Means, enabling analysis of larger datasets and models like LLMs.

Contribution

It introduces scalable clustering methods and metrics for latent concept discovery, significantly enhancing efficiency without sacrificing quality.

Findings

K-Means outperforms hierarchical clustering in efficiency

Metrics effectively assess concept quality

Scalable discovery applied to large language models

Abstract

Despite the revolution caused by deep NLP models, they remain black boxes, necessitating research to understand their decision-making processes. A recent work by Dalvi et al. (2022) carried out representation analysis through the lens of clustering latent spaces within pre-trained models (PLMs), but that approach is limited to small scale due to the high cost of running Agglomerative hierarchical clustering. This paper studies clustering algorithms in order to scale the discovery of encoded concepts in PLM representations to larger datasets and models. We propose metrics for assessing the quality of discovered latent concepts and use them to compare the studied clustering algorithms. We found that K-Means-based concept discovery significantly enhances efficiency while maintaining the quality of the obtained concepts. Furthermore, we demonstrate the practicality of this newfound efficiency by scaling latent concept discovery to LLMs and phrasal concepts.