Comparing Task-Specific Ensembles to General Language Models in Distinguishing Elaborated and Compact Language Codes within CoCA-CoLA