EUVIMEDEUVIMED entdecken
Crossref

Decoupling Knowledge and Reasoning in LLMs: An Exploration Using Cognitive Dual-System Theory

Mutian Yang, Jiandong Gao, Ji Wu

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (40) · 34268-34276

DOI: 10.1609/aaai.v40i40.40723 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Mutian Yang und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

While large language models (LLMs) leverage both knowledge and reasoning during inference, the capacity to distinguish between them plays a pivotal role in model analysis, interpretability, and development. Inspired by dual-system cognitive theory, we propose a cognition attribution framework to decouple the contribution of knowledge and reasoning. In particular, the cognition of LLMs is decomposed into two distinct yet complementary phases: knowledge retrieval (Phase 1) and reasoning adjustment (Phase 2). To separate these phases, LLMs are prompted to generate answers under two different cognitive modes, fast thinking and slow thinking, respectively. The performance under different cognitive modes is analyzed to quantify the contribution of knowledge and reasoning. This architecture is employed to 15 LLMs across 3 datasets. Results reveal: (1) reasoning adjustment is domain-specific, benefiting reasoning-intensive domains (e.g., mathematics, physics, and chemistry) and potentially imparing knowledge-intensive domains. (2) Parameter scaling improves both knowledge and reasoning, with knowledge improvements being more pronounced. Additionally, parameter scaling make LLMs reasoning significantly more prudent, while moderately more intelligent. (3) Knowledge primarily resides in lower network layers, while reasoning operates in higher layers. Our framework not only helps understand LLMs from a "decoupling" perspective, but also provides new insights into existing research, including scaling laws, hierarchical knowledge editing, and limitations of small-scale-LLM reasoning.

Zitieren

APA

Yang, M., Gao, J., & Wu, J. (2026). Decoupling Knowledge and Reasoning in LLMs: An Exploration Using Cognitive Dual-System Theory. Proceedings of the AAAI Conference on Artificial Intelligence, 40(40), 34268-34276. https://doi.org/10.1609/aaai.v40i40.40723

Vancouver

Yang M, Gao J, Wu J. Decoupling Knowledge and Reasoning in LLMs: An Exploration Using Cognitive Dual-System Theory. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(40):34268-34276. doi:10.1609/aaai.v40i40.40723

BibTeX
@article{Yang2026,
  author = {Yang, Mutian and Gao, Jiandong and Wu, Ji},
  title = {Decoupling Knowledge and Reasoning in LLMs: An Exploration Using Cognitive Dual-System Theory},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {40},
  pages = {34268--34276},
  doi = {10.1609/aaai.v40i40.40723}
}
Kostenlos starten →