Exploiting Hard Samples for Stealthy Backdoor Attacks on Large Language Models

Diqun Yan, Rangding Wang. Exploiting Hard Samples for Stealthy Backdoor Attacks on Large Language Models. In Xiaoliang Wang, Xiaohong Jiang, Noel Crespi, Baoliu Ye, editors, Network and Parallel Computing - 21st IFIP WG 10.3 International Conference, NPC 2025, Nha Trang, Vietnam, November 14-16, 2025, Proceedings, Part II. Volume 16306 of Lecture Notes in Computer Science, pages 203-214, Springer, 2025. [doi]

Abstract

Abstract is missing.