Fun-tuning: Characterizing the Vulnerability of Proprietary LLMs to Optimization-based Prompt Injection Attacks via the Fine-Tuning Interface

Name: Fun-tuning: Characterizing the Vulnerability of Proprietary LLMs to Optimization-based Prompt Injection Attacks via the Fine-Tuning Interface
Duration: 15 min
Description: Talk from IEEE Symposium on Security and Privacy 2025.

Andrey Labunets, Nishit V. Pandya, Ashish Hooda, Xiaohan Fu, Earlence Fernandes

IEEE Symposium on Security and Privacy 2025 · Day 1 · LLM Security