Investigating representations of verb bias in neural language models
Robert Hawkins, Takateru Yamakoshi, Thomas Griffiths, Adele Goldberg
Abstract
Languages typically provide more than one grammatical construction to express certain types of messages. A speaker’s choice of construction is known to depend on multiple factors, including the choice of main verb – a phenomenon known as verb bias. Here we introduce DAIS, a large benchmark dataset containing 50K human judgments for 5K distinct sentence pairs in the English dative alternation. This dataset includes 200 unique verbs and systematically varies the definiteness and length of arguments. We use this dataset, as well as an existing corpus of naturally occurring data, to evaluate how well recent neural language models capture human preferences. Results show that larger models perform better than smaller models, and transformer architectures (e.g. GPT-2) tend to out-perform recurrent architectures (e.g. LSTMs) even under comparable parameter and training settings. Additional analyses of internal feature representations suggest that transformers may better integrate specific lexical information with grammatical constructions.- Anthology ID:
- 2020.emnlp-main.376
- Volume:
- Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)
- Month:
- November
- Year:
- 2020
- Address:
- Online
- Editors:
- Bonnie Webber, Trevor Cohn, Yulan He, Yang Liu
- Venue:
- EMNLP
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 4653–4663
- Language:
- URL:
- https://aclanthology.org/2020.emnlp-main.376
- DOI:
- 10.18653/v1/2020.emnlp-main.376
- Cite (ACL):
- Robert Hawkins, Takateru Yamakoshi, Thomas Griffiths, and Adele Goldberg. 2020. Investigating representations of verb bias in neural language models. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 4653–4663, Online. Association for Computational Linguistics.
- Cite (Informal):
- Investigating representations of verb bias in neural language models (Hawkins et al., EMNLP 2020)
- PDF:
- https://preview.aclanthology.org/naacl24-info/2020.emnlp-main.376.pdf
- Code
- taka-yamakoshi/neural_constructions
- Data
- DAIS