EdgePrompt introduces a novel cloud-edge collaborative framework that addresses the critical security and latency challenges of deploying Large Language Models in 6G networks.

• Implements a hierarchical attention splicing mechanism to distribute computational load
• Incorporates privacy-preserving strategies by isolating sensitive information
• Reduces data leakage risks while maintaining high performance
• Enables edge-based processing for time-sensitive applications

This research is crucial for security professionals as it demonstrates how to deploy AI capabilities in next-generation networks while maintaining data sovereignty and reducing attack surfaces through distributed processing.

EdgePrompt: A Distributed Key-Value Inference Framework for LLMs in 6G Networks