From BERT‘s Point of View: Revealing the Prevailing Contextual Differences

Carolin M. Schuster, Simon Hegelich


Abstract
Though successfully applied in research and industry large pretrained language models of the BERT family are not yet fully understood. While much research in the field of BERTology has tested whether specific knowledge can be extracted from layer activations, we invert the popular probing design to analyze the prevailing differences and clusters in BERT’s high dimensional space. By extracting coarse features from masked token representations and predicting them by probing models with access to only partial information we can apprehend the variation from ‘BERT’s point of view’. By applying our new methodology to different datasets we show how much the differences can be described by syntax but further how they are to a great extent shaped by the most simple positional information.
Anthology ID:
2022.findings-acl.89
Volume:
Findings of the Association for Computational Linguistics: ACL 2022
Month:
May
Year:
2022
Address:
Dublin, Ireland
Editors:
Smaranda Muresan, Preslav Nakov, Aline Villavicencio
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1120–1138
Language:
URL:
https://aclanthology.org/2022.findings-acl.89
DOI:
10.18653/v1/2022.findings-acl.89
Bibkey:
Cite (ACL):
Carolin M. Schuster and Simon Hegelich. 2022. From BERT‘s Point of View: Revealing the Prevailing Contextual Differences. In Findings of the Association for Computational Linguistics: ACL 2022, pages 1120–1138, Dublin, Ireland. Association for Computational Linguistics.
Cite (Informal):
From BERT‘s Point of View: Revealing the Prevailing Contextual Differences (Schuster & Hegelich, Findings 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.findings-acl.89.pdf
Software:
 2022.findings-acl.89.software.zip