| 1 |
Probing for the Usage of Grammatical Number ...
|
|
|
|
Abstract:
A central quest of probing is to uncover how pre-trained models encode a linguistic property within their representations. An encoding, however, might be spurious-i.e., the model might not rely on it when making predictions. In this paper, we try to find encodings that the model actually uses, introducing a usage-based probing setup. We first choose a behavioral task which cannot be solved without using the linguistic property. Then, we attempt to remove the property by intervening on the model's representations. We contend that, if an encoding is used by the model, its removal should harm the performance on the chosen behavioral task. As a case study, we focus on how BERT encodes grammatical number, and on how it uses this encoding to solve the number agreement task. Experimentally, we find that BERT relies on a linear encoding of grammatical number to produce the correct behavioral output. We also find that BERT uses a separate encoding of grammatical number for nouns and verbs. Finally, we identify in ... : ACL 2022 (Main Conference) The discussion section had been inadvertently removed before the article was published on arxiv ...
|
|
Keyword:
Computation and Language cs.CL; FOS Computer and information sciences
|
|
URL: https://arxiv.org/abs/2204.08831
https://dx.doi.org/10.48550/arxiv.2204.08831
|
|
BASE
|
|
Hide details
|
|
|
| 9 |
Finding Concept-specific Biases in Form--Meaning Associations ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 10 |
Searching for Search Errors in Neural Morphological Inflection ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 11 |
Applying the Transformer to Character-level Transduction ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 12 |
Quantifying Gender Bias Towards Politicians in Cross-Lingual Language Models ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 17 |
Examining the Inductive Bias of Neural Language Models with Artificial Languages ...
|
|
|
|
BASE
|
|
Show details
|
|
|
|
|
