Detecting Label Errors in Token Classification Data

Wei-Chen Wang, Jonas Mueller

Mislabeled examples are a common issue in real-world data, particularly for tasks like token classification where many labels must be chosen on a fine-grained basis. Here we consid…