Teaching Broad Reasoning Skills via Decomposition-Guided Contexts. (arXiv:2205.12496v1 [cs.CL])

Question-answering datasets require a broad set of reasoning skills. We show
how to use question decompositions to teach language models these broad
reasoning skills in a robust fashion. Specifically, we use widely available
QDMR representations to programmatically create synthetic contexts for real
questions in six multihop reasoning datasets. These contexts are carefully
designed to avoid common reasoning shortcuts prevalent in real contexts that
prevent models from learning the right skills. This results in a pretraining
dataset, named TeaBReaC, containing 525K …

arxiv reasoning skills teaching