metadata
license: cc-by-sa-3.0
This dataset is a deduplicated subset of the validation split of BoolQ, as used in the paper How Much Can We Forget about Data Contamination?. The deduplication was performed using this script.
The data fields are the same as in https://huggingface.co/datasets/google/boolq, with the additional "split-id" column that can be used to partition the benchmark questions into different subsets. The dataset can be used as a plug-in replacement for BoolQ if you want to work with the deduplicated benchmark questions.
For details on how the dataset was constructed, see Section A.4 and Table 2 in the paper.