This 1.5B-model is based on Qwen-1.5B-Instruct and it is trained following AgentStealth pipeline. Github repo is
https://github.com/tsinghua-fib-lab/AgentStealth.
Disclaimer & Ethical Usage
This model is released for research and educational purposes only. Its primary intent is to serve as a benchmark tool for testing and evaluating deanonymization defenses and improving privacy-preserving techniques.
Strict Prohibition of Misuse:
This model MUST NOT be used to conduct actual privacy attacks, de-anonymize real-world individuals, or extract sensitive information from any private datasets.
Users are strictly prohibited from using this model for any malicious activities that violate data protection laws (such as GDPR, PIPL, etc.) or individual privacy rights.
The authors shall not be held responsible for any misuse of this model by third parties. By downloading or using this model, you agree to abide by these ethical guidelines and use it solely for defensive research.
