Under the Hood: Using Diagnostic Classifiers to Investigate and Improve how Language Models Track Agreement Information

Open in new window