BOOSTING MATHEMATICAL REASONING OF LLMS

Summary from the TLDR AI newsletter

Researchers investigate the impact of data augmentation techniques on improving the math reasoning capabilities of large language models. They create a new dataset, AugGSM8K, by enhancing queries from an existing dataset and developing a model named MuggleMath.