There's the HANS dataset which is used to evaluate NLI models to check if they pick up surface heuristics: https://huggingface.co/datasets/hans https://huggingf

There's the HANS dataset which is used to evaluate NLI models to check if they pick up surface heuristics: https://huggingface.co/datasets/hans https://huggingface.co/datasets/hans https://arxiv.org/abs/1902.01007 https://arxiv.org/abs/1902.01007