Uncovering Hidden Consequences of Pre-training Objectives in Sequence-to-Sequence Models

Some variants of self-supervised denoising objectives for pre-training encoder-decoder language models have been reported to have a negligible impact on downstream performance. Yet the design of these pre-training objectives leads to behavioural differences that can be uncovered with specific manipulations. We reproduce a recently proposed zero-shot control method and find that it is only successful on a subset of models. To understand what causes the difference in its effectiveness, we perform a set of controlled experiments, varying only the pre-training objective, and find unexpected interactions between the pre-training method and downstream controllability of models after fine-tuning. Our results show that different pretraining objectives have consequences that may not be visible in standard downstream evaluation, but which should be taken into account when developing models with controllability in mind.