Cross-Domain and Cross-Language Irony Detection: The Impact of Bias on Models’ Generalization | Publicación