Life after BERT: What do Other Muppets Understand about Language?

Anonymous

Abstract: Pre-trained transformers are at the core of natural language processing today. However, the understanding of what model learns during pre-training is still limited. Existing model analysis works usually focus only on one or two model families at a time, overlooking the variety of existing architectures and pre-training objectives. In our work, we utilize the oLMpics benchmark and psycholinguistic probing datasets for a diverse set of 28 models including T5, BART, and ALBERT. Additionally, we adapt the oLMpics zero-shot setup for autoregressive models and evaluate GPT networks of different sizes. Our findings show that none of these models can resolve compositional questions in a zero-shot fashion, suggesting that this skill is not learnable using existing pre-training objectives. Additionally, we find that global model decisions such as architecture, directionality, size of the dataset, and pre-training objective are not predictive of a model's linguistic capabilities.