HypoNLI: Exploring the Artificial Patterns of Hypothesis-only Bias in Natural Language Inference

Open in new window