A collection of research on enhancing the reasoning capabilities of large language models, particularly over structured, symbolic, and tabular data.