<ul><li>Graph prompt tuning has emerged as a promising paradigm for transferring general graph knowledge from pre-trained models to downstream tasks.</li><li>A backdoor attack known as Krait has been introduced, which disguises benign graph prompts to evade detection.</li><li>Krait efficiently embeds triggers to a small fraction of training nodes, achieving high attack success rates without sacrificing clean accuracy.</li><li>The study analyzes how Krait can evade both classical and state-of-the-art defenses and provides insights for detecting and mitigating such attacks.</li></ul>

Krait: A Backdoor Attack Against Graph Prompt Tuning

Discover more