Verb Conjugation in Transformers Is Determined by Linear Encodings of Subject Number

Findings of the Association for Computational Linguistics: EMNLP 2023

Venue: EMNLP
Type: Conference
Causal Intervention
Probing
Interpretability
BERTology
Authors
Affiliations

Sophie Hao

New York University

Tal Linzen

New York University

Google

Published

December 6, 2023

Abstract
Deep architectures such as Transformers are sometimes criticized for having uninterpretable “black-box” representations. We use causal intervention analysis to show that, in fact, some linguistic features are represented in a linear, interpretable format. Specifically, we show that BERT’s ability to conjugate verbs relies on a linear encoding of subject number that can be manipulated with predictable effects on conjugation accuracy. This encoding is found in the subject position at the first layer and the verb position at the last layer, but distributed across positions at middle layers, particularly when there are multiple cues to subject number.