The dataset viewer is not available for this split.
Rows from parquet row groups are too big to be read: 3.53 GiB (max=286.10 MiB)
Error code: TooBigContentError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
We present SpuriVerse, a multimodal benchmark comprising 124 distinct spurious correlations extracted from real-world datasets, each containing 11 examples --- 1 from the original dataset, referred to as the anchor and 10 synthetic examples, referred as the spurious group --- for a total of 1364 instances.
- Downloads last month
- 13