While language models are powerful and versatile, they often fail to address highly complex problems. This is because solving complex problems requires deliberate thinking, which has been only minimally guided during training. In this paper, we propose a new method called Cumulative Reasoning (CR), which employs language models in a cumulative and iterative manner to emulate human thought processes. By decomposing tasks into smaller components, CR streamlines the problem-solving process, rendering it both more manageable and effective. For logical inference tasks, CR consistently outperforms existing methods with an improvement up to 9.3%, and achieves the astonishing accuracy of 98.04% on the curated FOLIO wiki dataset. In the context of the Game of 24, CR achieves an accuracy of 94%, which signifies a substantial enhancement of 20% over the previous state-of-the-art method.

翻译：尽管语言模型功能强大且用途广泛，但它们在处理高度复杂问题时往往表现不佳。这是因为解决复杂问题需要深思熟虑的思考，而在训练过程中对这种思考的引导十分有限。本文提出一种名为累积推理（Cumulative Reasoning, CR）的新方法，该方法通过累积和迭代的方式运用语言模型，以模拟人类的思维过程。通过将任务分解为更小的组成部分，CR简化了问题解决流程，使其更易于管理和高效。在逻辑推理任务中，CR始终优于现有方法，性能提升最高达9.3%，并在精心策划的FOLIO wiki数据集上实现了98.04%的惊人准确率。在“24点游戏”的背景下，CR达到了94%的准确率，相较于先前的最先进方法显著提升了20%。