Measuring Distributional Shifts in Text: The Advantage of Language Model-Based Embeddings

Open in new window