Towards LLM-Enhanced Android Taint Analysis
Whether off-the-shelf Large Language Models (LLMs) can effectively reason about taint flows in Android apps is investigated, and preliminary findings suggest that LLM reasoning may effectively complement traditional static taint analysis.
Nicholas Miazzo, Marco Alecci, Jordan Samhi et al.
· 0 citations