Under the Hood: Using Diagnostic Classifiers to Investigate and Improve how Language Models Track Agreement Information (original) (raw)

Causal Analysis of Syntactic Agreement Mechanisms in Neural Language Models

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)

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 study applies causal mediation analysis to pre-trained neural language models. We investigate the magnitude of models' preferences for grammatical inflections, as well as whether neurons process subject-verb agreement similarly across sentences with different syntactic structures. We uncover similarities and differences across architectures and model sizesnotably, that larger models do not necessarily learn stronger preferences. We also observe two distinct mechanisms for producing subject-verb agreement depending on the syntactic structure of the input sentence. Finally, we find that language models rely on similar sets of neurons when given sentences with similar syntactic structure. * Equal contribution. † Work done while visiting Google Research. ‡ Supported by the Viterbi Fellowship in the Center for Computer Engineering at the Technion.

Can RNNs trained on harder subject-verb agreement instances still perform well on easier ones?

ArXiv, 2021

The main subject and the associated verb in English must agree in grammatical number as per the Subject-Verb Agreement (SVA) phenomenon. It has been found that the presence of a noun between the verb and the main subject, whose grammatical number is opposite to that of the main subject, can cause speakers to produce a verb that agrees with the intervening noun rather than the main noun; the former thus acts as an agreement attractor. Such attractors have also been shown to pose a challenge for RNN models without explicit hierarchical bias to perform well on SVA tasks. Previous work suggests that syntactic cues in the input can aid such models to choose hierarchical rules over linear rules for number agreement. In this work, we investigate the effects of the choice of training data, training algorithm, and architecture on hierarchical generalization. We observe that the models under consideration fail to perform well on sentences with no agreement attractor when trained solely on nat...

Does BERT really agree ? Fine-grained Analysis of Lexical Dependence on a Syntactic Task

Findings of the Association for Computational Linguistics: ACL 2022

Although transformer-based Neural Language Models demonstrate impressive performance on a variety of tasks, their generalization abilities are not well understood. They have been shown to perform strongly on subject-verb number agreement in a wide array of settings, suggesting that they learned to track syntactic dependencies during their training even without explicit supervision. In this paper, we examine the extent to which BERT is able to perform lexically-independent subject-verb number agreement (NA) on targeted syntactic templates. To do so, we disrupt the lexical patterns found in naturally occurring stimuli for each targeted structure in a novel fine-grained analysis of BERT's behavior. Our results on nonce sentences suggest that the model generalizes well for simple templates, but fails to perform lexically-independent syntactic generalization when as little as one attractor is present.

Exploring Processing of Nested Dependencies in Neural-Network Language Models and Humans

ArXiv, 2020

Recursive processing in sentence comprehension is considered a hallmark of human linguistic abilities. However, its underlying neural mechanisms remain largely unknown. We studied whether a recurrent neural network with Long Short-Term Memory units can mimic a central aspect of human sentence processing, namely the handling of long-distance agreement dependencies. Although the network was solely trained to predict the next word in a large corpus, analysis showed the emergence of a small set of specialized units that successfully handled local and long-distance syntactic agreement for grammatical number. However, simulations showed that this mechanism does not support full recursion and fails with some long-range embedded dependencies. We tested the model's predictions in a behavioral experiment where humans detected violations in number agreement in sentences with systematic variations in the singular/plural status of multiple nouns, with or without embedding. Human and model er...

Subject Verb Agreement Error Patterns in Meaningless Sentences: Humans vs. BERT

2022

Both humans and neural language models are able to perform subject-verb number agreement (SVA). In principle, semantics shouldn't interfere with this task, which only requires syntactic knowledge. In this work we test whether meaning interferes with this type of agreement in English in syntactic structures of various complexities. To do so, we generate both semantically well-formed and nonsensical items. We compare the performance of BERT-base to that of humans, obtained with a psycholinguistic online crowdsourcing experiment. We find that BERT and humans are both sensitive to our semantic manipulation: They fail more often when presented with nonsensical items, especially when their syntactic structure features an attractor (a noun phrase between the subject and the verb that has not the same number as the subject). We also find that the effect of meaningfulness on SVA errors is stronger for BERT than for humans, showing higher lexical sensitivity of the former on this task.

Can RNNs learn Recursive Nested Subject-Verb Agreements?

ArXiv, 2021

One of the fundamental principles of contemporary linguistics states that language processing requires the ability to extract recursively nested tree structures. However, it remains unclear whether and how this code could be implemented in neural circuits. Recent advances in Recurrent Neural Networks (RNNs), which achieve near-human performance in some language tasks, provide a compelling model to address such questions. Here, we present a new framework to study recursive processing in RNNs, using subject-verb agreement as a probe into the representations of the neural network. We trained six distinct types of RNNs on a simplified probabilistic context-free grammar designed to independently manipulate the length of a sentence and the depth of its syntactic tree. All RNNs generalized to subject-verb dependencies longer than those seen during training. However, none systematically generalized to deeper tree structures, even those with a structural bias towards learning nested tree (i....

Targeted Syntactic Evaluation of Language Models

Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, 2018

We present a dataset for evaluating the grammaticality of the predictions of a language model. We automatically construct a large number of minimally different pairs of English sentences, each consisting of a grammatical and an ungrammatical sentence. The sentence pairs represent different variations of structure-sensitive phenomena: subject-verb agreement, reflexive anaphora and negative polarity items. We expect a language model to assign a higher probability to the grammatical sentence than the ungrammatical one. In an experiment using this data set, an LSTM language model performed poorly on many of the constructions. Multi-task training with a syntactic objective (CCG supertagging) improved the LSTM's accuracy, but a large gap remained between its performance and the accuracy of human participants recruited online. This suggests that there is considerable room for improvement over LSTMs in capturing syntax in a language model.

The emergence of number and syntax units in

Proceedings of the 2019 Conference of the North, 2019

Recent work has shown that LSTMs trained on a generic language modeling objective capture syntax-sensitive generalizations such as longdistance number agreement. We have however no mechanistic understanding of how they accomplish this remarkable feat. Some have conjectured it depends on heuristics that do not truly take hierarchical structure into account. We present here a detailed study of the inner mechanics of number tracking in LSTMs at the single neuron level. We discover that longdistance number information is largely managed by two "number units". Importantly, the behaviour of these units is partially controlled by other units independently shown to track syntactic structure. We conclude that LSTMs are, to some extent, implementing genuinely syntactic processing mechanisms, paving the way to a more general understanding of grammatical encoding in LSTMs.

Making Syntax of Sense: Number Agreement in Sentence Production

Psychological Review, 2005

Grammatical agreement flags the parts of sentences that belong together regardless of whether the parts appear together. In English, the major agreement controller is the sentence subject, the major agreement targets are verbs and pronouns, and the major agreement category is number. The authors expand an account of number agreement whose tenets are that pronouns acquire number lexically, whereas verbs acquire it syntactically but with similar contributions from number meaning and from the number morphology of agreement controllers. These tenets were instantiated in a model using existing verb agreement data. The model was then fit to a new, more extensive set of verb data and tested with a parallel set of pronoun data. The theory was supported by the model's outcomes. The results have implications for the integration of words and structures, for the workings of agreement categories, and for the nature of the transition from thought to language.

Serial Recall Effects in Neural Language Modeling

Proceedings of the 2019 Conference of the North

Serial recall experiments study the ability of humans to recall words in the order in which they occurred. The following serial recall effects are generally investigated in studies with humans: word length and frequency, primacy and recency, semantic confusion, repetition, and transposition effects. In this research, we investigate LSTM language models in the context of these serial recall effects. Our work provides a framework to better understand and analyze neural language models and opens a new window to develop accurate language models.