View of Task-Aware Regularized Continual Fine-Tuning for Mitigating Catastrophic Forgetting in Large Language Models | Journal of Computer Technology and Software