Bo-Le Ma

The Illusion of Power Capping in LLM Decode: A Phase-Aware Energy Characterisation Across Attention Architectures

It is found that, once prefill is measured without prefix caching, the recurrent architectures pay no prefill penalty: their linear-time mixing undercuts quadratic attention on long prompts, so their efficient decode adds to the prefill saving and lowers total request energy from the first output token at large batch.