Inversely Eliciting Numerical Reasoning in Language Models via Solving Linear Systems

Recent language models have struggled to generalize to a large range of numbers in numerical reasoning.In this paper, we propose a novel method that leverages simple numbers as anchors to characterize the implicitly inferred arithmetic expressions from language models, and then explicitly applies the expressions to original numbers to get the answers.Experimental results on several numerical reasoning benchmarks demonstrate that our approach is highly effective.More importantly, our approach works in the inference phase without extra model training, making it highly portable and achieving significant and consistent performance benefits across a variety of language models in zero-shot, few-shot, and fine-tuning scenarios.