This deliverable details the PRIVATEER approach towards machine learning models that enable automatic security analytics at and across base stations in 6G networks, while maintaining robustness and privacy, providing efficient training and inference, and ensuring that we can explain why a model classifies traffic as benign or malicious.