- Sirdaryo viloyati Guliston shahar
- Dushanba-Shanba: 9.00 - 18.00
Azizbek Xaitbayev Pirnazarovich
10.5281/zenodo.22939915
Annotatsiya
Katta til modellari va Generativ sun’iy intellekt (GenAI) xizmatlarining korporativ hamda individual jarayonlarga jadal integratsiyalashuvi yangi turdagi kiberxavfsizlik tahdidlarini keltirib chiqarmoqda. Ular orasida eng xavflisi Prompt Injection xurujlari bo‘lib, ular modelning mantiqiy ko‘rsatmalarini buzish va tizim ichidagi maxfiy ma’lumotlarni oshkor qilishga qaratilgan. Ushbu tezisda Prompt Injection xurujlarining mexanizmlari, ularning ma’lumotlar sizib chiqishiga ta’siri hamda ushbu xavflarni kamaytirishga qaratilgan ko‘p qatlamli amaliy va texnik yechimlar tahlil qilinadi.
Kalit so'zlar:
Generativ AI tizimlarida ma’lumotni qayta ishlashning an’anaviy dasturiy ta’minotdan asosiy farqi shundaki, tizim yo‘riqnomasi va foydalanuvchi ma’lumoti bitta matnli oqimda modelga uzatiladi. Bu holat arxitektura darajasidagi zaiflikni yuzaga keltiradi.
Foydalanilgan adabiyotlar
1. Greshake, K., Abdelnabi, S., Mishra, S., Endres, C., Holz, T., & Fritz, M. (2023). Not what you’ve signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection. Proceedings of the 16th ACM Workshop on Artificial Intelligence and Security (AISEC ’23), 79–90. 2. OWASP Foundation. (2023). OWASP Top 10 for Large Language Model Applications (v1.1). Open Web Application Security Project. 3. Inan, H., Upasani, K., et al. (2023). Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations. arXiv preprint arXiv:2312.06674. 4. Yao, Y., Duan, J., et al. (2024). A Survey on Large Language Model (LLM) Security and Privacy: Vulnerabilities, Attacks, and Defenses. IEEE TKDE, 1–20.