[1]
Y. Yan, “S^3cMath: Spontaneous Step-Level Self-Correction Makes Large Language Models Better Mathematical Reasoners”, AAAI, vol. 39, no. 24, pp. 25588–25596, Apr. 2025.