A Study on How Attention Scores in the BERT Model Are Aware of Lexical Categories in Syntactic and Semantic Tasks on the GLUE Benchmark

Dongjun Jang, Sungjoo Byun, Hyopil Shin

Abstract
This study examines whether the attention scores between tokens in the BERT model significantly vary based on lexical categories during the fine-tuning process for downstream tasks. Drawing inspiration from the notion that in human language processing, syntactic and semantic information is parsed differently, we categorize tokens in sentences according to their lexical categories and focus on changes in attention scores among these categories. Our hypothesis posits that in downstream tasks that prioritize semantic information, attention scores centered on content words are enhanced, while in cases emphasizing syntactic information, attention scores centered on function words are intensified. Through experimentation conducted on six tasks from the GLUE benchmark dataset, we substantiate our hypothesis regarding the fine-tuning process. Furthermore, our additional investigations reveal the presence of BERT layers that consistently assign more bias to specific lexical categories, irrespective of the task, highlighting the existence of task-agnostic lexical category preferences.
Anthology ID:
2024.lrec-main.148
Volume:
Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)
Month:
May
Year:
2024
Address:
Torino, Italia
Editors:
Nicoletta Calzolari, Min-Yen Kan, Veronique Hoste, Alessandro Lenci, Sakriani Sakti, Nianwen Xue
Venues:
LREC | COLING
SIG:
Publisher:
ELRA and ICCL
Note:
Pages:
1684–1689
Language:
URL:
https://aclanthology.org/2024.lrec-main.148
DOI:
Bibkey:
Cite (ACL):
Dongjun Jang, Sungjoo Byun, and Hyopil Shin. 2024. A Study on How Attention Scores in the BERT Model Are Aware of Lexical Categories in Syntactic and Semantic Tasks on the GLUE Benchmark. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pages 1684–1689, Torino, Italia. ELRA and ICCL.
Cite (Informal):
A Study on How Attention Scores in the BERT Model Are Aware of Lexical Categories in Syntactic and Semantic Tasks on the GLUE Benchmark (Jang et al., LREC-COLING 2024)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-2/2024.lrec-main.148.pdf
PDF Search