This paper utilizes light-weight adapters that can be easily integrated with PLMs and serve as a repository for facts learned from different KBs and introduces Knowledge Probing using Response Selection (KPRS) – a probe designed specifically for TOD models.
We introduce the novel Wino-X benchmark to investigate whether translation models can perform coreference resolution that requires commonsense knowledge and whether multilingual language models are capable of commonsense reasoning across multiple languages. Our findings indicate that models are prone to biases and often fail to identify disambiguating information.
We introduce a method for the prediction of disambiguation errors based on statistical data properties, and develop a simple adversarial attack strategy that minimally perturbs sentences in order to elicit disambiguation errors to further probe the robustness of translation models.