Silver Dataset for Biomedical Multi-hop Question Decomposition Mode Classification

This is a silver dataset of 1,062 synthetic biomedical multi-hop questions specifically designed to train language models (such as BioBERT) for decomposition mode classification. The dataset includes the questions and their corresponding mode labels, and does not include the answers to the questions. It consists of two splits: a training set and a test set. The labels are encoded as follows: 0 for Direct, 1 for Sequential, 2 for Parallel.