ZHANG, R. Privacy-Oriented Text Generation in LLMs via Selective Fine-Tuning and Semantic Attention Masks . Journal of Computer Technology and Software, v. 4, n. 8, 30 Aug.2025.