Pre-Calc: Learning to Use the Calculator Improves Numeracy in Language Models (2404.14355v3)

Abstract: Quantitative and numerical comprehension in language is an important task in many fields like education and finance, but still remains a challenging task for LLMs. While tool and calculator usage has shown to be helpful to improve mathematical reasoning in large pretrained decoder-only LLMs, this remains unexplored for smaller LLMs with encoders. In this paper, we propose Pre-Calc, a simple pre-finetuning objective of learning to use the calculator for both encoder-only and encoder-decoder architectures, formulated as a discriminative and generative task respectively. We pre-train BERT and RoBERTa for discriminative calculator use and Flan-T5 for generative calculator use on the MAWPS, SVAMP, and AsDiv-A datasets, which improves performance on downstream tasks that require numerical understanding. Our code and data are available at https://github.com/calc-cmu/pre-calc.