CAMA: Enhancing Mathematical Reasoning in Large Language Models with Causal Knowledge

Lei Zan, Keli Zhang, Ruichu Cai, Lujia Pan · 2025

Read CAMA: Enhancing Mathematical Reasoning in Large Language Models with Causal Knowledge by Lei Zan, Keli Zhang, Ruichu Cai, Lujia Pan (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-02583.txt.

Description

Lei Zan and co-authors gave language models a causal map for math in two thousand twenty-five. Models still stumble on complex math reasoning. The failures trace to deep structure linking solution steps. Their framework, CAMA, works in two stages. First it builds a graph of math knowledge points. Causal links between those points are learned from examples.

Example question and solution pairs feed the graph building. A language model's prior knowledge seeds the process. Discovery algorithms then sharpen the causal edges. Feedback from chosen examples refines the graph further. At solving time, the graph guides the model's reasoning. Reusable structure replaces one-off memorized tricks. The authors report stronger performance on hard problems.

Catalog record

CREATORLei Zan, Keli Zhang, Ruichu Cai, Lujia Pan
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE57 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-02583

Browse the full library · About the Tower of Babel