Causal Analysis of Syntactic Agreement Neurons in Multilingual Language Models
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.
Code (1)
Tasks
counterfactualSimilar Papers 제목 키워드 기반
Causal Analysis of Syntactic Agreement Mechanisms in Neural Language Models
Targeted syntactic evaluations have demonstrated the ability of language models to perform subject-verb agreement given difficult contexts. To elucidate the mechanisms by which the models accomplish this behavior, this s…
SentenceShared Circuits for Shared Grammar: Tracing Subject-Verb Agreement Across Languages
Multilingual large language models often generalize across languages, and prior work suggests that their internal mechanisms can overlap cross-lingually. It remains unclear, however, when such sharing emerges and whether…
How Distributed are Distributed Representations? An Observation on the Locality of Syntactic Information in Verb Agreement Tasks
This work addresses the question of the localization of syntactic information encoded in the transformers representations. We tackle this question from two perspectives, considering the object-past participle agreement i…
feature selectionSentenceHow Distributed are Distributed Representations? An Observation on the Locality of Syntactic Information in Verb Agreement Tasks
This work addresses the question of the localization of the syntactic information encoding in the Transformers representations. We tackle this question from two perspectives, considering the object-past participle agreem…
feature selectionSentenceControlled Evaluation of Syntactic Knowledge in Multilingual Language Models
Language models (LMs) are capable of acquiring elements of human-like syntactic knowledge. Targeted syntactic evaluation tests have been employed to measure how well they form generalizations about syntactic phenomena in…
Diversity