Transformer-based language models are known to display anisotropic behav...
An ongoing debate in the NLG community concerns the best way to evaluate...
In this paper we investigate the linguistic knowledge learned by a Neura...
In the last few years, pre-trained neural architectures have provided
im...
share