CoTSRF: Utilize Chain of Thought as Stealthy and Robust Fingerprint of Large Language Models

Open in new window