Sim-LLM: Optimizing LLM Inference at the Edge through Inter-Task KV Reuse

KV cache technology, by storing key-value pairs, helps reduce the computational overhead incurred by (LLMs).