Task-Aware Regularized Continual Fine-Tuning for Mitigating Catastrophic Forgetting in Large Language Models. Journal of Computer Technology and Software, v. 4, n. 7, 30 July2025.