DeepAI AI Chat
Log In Sign Up

The emergence of number and syntax units in LSTM language models

by   Yair Lakretz, et al.
University of Amsterdam

Recent work has shown that LSTMs trained on a generic language modeling objective capture syntax-sensitive generalizations such as long-distance 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 long-distance 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.


page 1

page 2

page 3

page 4


Do LSTMs See Gender? Probing the Ability of LSTMs to Learn Abstract Syntactic Rules

LSTMs trained on next-word prediction can accurately perform linguistic ...

Modeling German Verb Argument Structures: LSTMs vs. Humans

LSTMs have proven very successful at language modeling. However, it rema...

Assessing the Unitary RNN as an End-to-End Compositional Model of Syntax

We show that both an LSTM and a unitary-evolution recurrent neural netwo...

LSTMs Compose (and Learn) Bottom-Up

Recent work in NLP shows that LSTM language models capture hierarchical ...

Colorless green recurrent networks dream hierarchically

Recurrent neural networks (RNNs) have achieved impressive results in a v...

Predictive Representation Learning for Language Modeling

To effectively perform the task of next-word prediction, long short-term...

Code Repositories



view repo