LLMs (Large Language Models) often struggle to balance compositionality with knowledgeability, a challenge we define as the 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.
Blogger's Review: The introduction of CPP marks a significant advancement in the reasoning capabilities of large language models, especially in fields demanding precise knowledge. This approach not only improves model performance but also paves the way for new research avenues. The integration of compositionality and knowledge-based frameworks will drive the realization of more practical applications.