让大模型推理更准：把抽象命题变具体，提升医疗和数学任务表现
Concretized Proposition Prompting Resolves Composition-Knowledge Dichotomy in Large Language Models
- 将问题中的命题具体化，引导模型聚焦真实事实
- 医疗推理准确率显著提升，数学推理也保持竞争力
- 适配多种模型规模，可作为通用推理增强框架
大语言模型常难以兼顾组合性与知识性，我们称之为组合-知识二元困境。为此提出具象化命题提示（CPP），通过显式具象化与问题相关的命题来解决该问题。实验表明，CPP显著提升推理性能，尤其在对精确知识要求高的医疗基准上表现突出，同时在以演绎推理为主的数学基准上也具备竞争力。额外实验显示，CPP可扩展至不同基础模型与参数规模，是一种能弥合组合型与知识型方法差距的根本性范式，从而为逻辑有序、事实可靠的推理提供坚实基础。
原文摘要 · Abstract (English)
LLMs often struggle to balance compositionality with knowledgeability, a challenge we define as Composition-Knowledge Dichotomy. To address this, we propose Concretized Proposition Prompting (CPP), a framework that explicitly concretizes propositions relevant to questions. The results demonstrate that CPP significantly enhances reasoning performance, particularly in medical benchmarks where precise knowledge is paramount, while being competitive on math benchmarks where deductive reasoning is prioritized. Additional experiments reveal that CPP is scalable to various foundation models and parameter sizes, being a fundamental paradigm that bridges the gap between composition- and knowledge-based approaches. Consequently, CPP resolves the composition-knowledge dichotomy by providing a solid foundation for logically organized and factually grounded reasoning.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。