Abstract
Structural probing work has found evidence for latent syntactic information in pre-trained language models. However, much of this analysis has focused on monolingual models, and analyses of multilingual models have employed correlational methods that are confounded by the choice of probing tasks. In this study, we causally probe multilingual language models (XGLM and multilingual BERT) as well as monolingual BERT-based models across various languages; we do this by performing counterfactual perturbations on neuron activations and observing the effect on models’ subject-verb agreement probabilities. We observe where in the model and to what extent syntactic agreement is encoded in each language. We find significant neuron overlap across languages in autoregressive multilingual language models, but not masked language models. We also find two distinct layer-wise effect patterns and two distinct sets of neurons used for syntactic agreement, depending on whether the subject and verb are separated by other tokens. Finally, we find that behavioral analyses of language models are likely underestimating how sensitive masked language models are to syntactic information.- Anthology ID:
- 2022.conll-1.8
- Volume:
- Proceedings of the 26th Conference on Computational Natural Language Learning (CoNLL)
- Month:
- December
- Year:
- 2022
- Address:
- Abu Dhabi, United Arab Emirates (Hybrid)
- Editors:
- Antske Fokkens, Vivek Srikumar
- Venue:
- CoNLL
- SIG:
- SIGNLL
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 95–109
- Language:
- URL:
- https://aclanthology.org/2022.conll-1.8
- DOI:
- 10.18653/v1/2022.conll-1.8
- Cite (ACL):
- Aaron Mueller, Yu Xia, and Tal Linzen. 2022. Causal Analysis of Syntactic Agreement Neurons in Multilingual Language Models. In Proceedings of the 26th Conference on Computational Natural Language Learning (CoNLL), pages 95–109, Abu Dhabi, United Arab Emirates (Hybrid). Association for Computational Linguistics.
- Cite (Informal):
- Causal Analysis of Syntactic Agreement Neurons in Multilingual Language Models (Mueller et al., CoNLL 2022)
- PDF:
- https://preview.aclanthology.org/dois-2013-emnlp/2022.conll-1.8.pdf