Discovering Useful Sentence Representations from Large Pretrained Language Models

Abstract

Despite the extensive success of pretrained language models as encoders for building NLP systems, they haven't seen prominence as decoders for sequence generation tasks. We explore the question of whether these models can be adapted to be used as universal decoders. To be considered "universal," a decoder must have an implicit representation for any target sentence $s$, such that it can recover that sentence exactly when conditioned on its representation. For large transformer-based language models trained on vast amounts of English text, we investigate whether such representations can be easily discovered using standard optimization methods. We present and compare three representation injection techniques for transformer-based models and three accompanying methods which map sentences to and from this representation space. Experiments show that not only do representations exist for sentences from a variety of genres. More importantly, without needing complex optimization algorithms, our methods recover these sentences almost perfectly without fine-tuning the underlying language model at all.

13 pages, 11 figures, 2 tables

References in corpus (8)

• MASS: Masked Sequence to Sequence Pre-training for Language Generation
• On Using Monolingual Corpora in Neural Machine Translation
• Layer Normalization
• Pointer Sentinel Mixture Models
• Chameleons in imagined conversations: A new approach to understanding coordination of linguistic style in dialogs
• KERMIT: Generative Insertion-Based Modeling for Sequences
• A Systematic Assessment of Syntactic Generalization in Neural Language Models
• Pre-trained Language Model Representations for Language Generation