Search papers, labs, and topics across Lattice.
This study investigates how variations in prompt design, specifically cognitive load and phrasing patterns, influence energy consumption during on-device inference of large language models (LLMs). By conducting a comprehensive empirical analysis that includes phase-level profiling to distinguish between prefill and decode energy costs, the authors reveal that cognitive load significantly impacts energy expenditure per token, while phrasing patterns primarily affect energy through token utilization. The findings underscore the necessity of model-aware prompt design to optimize energy efficiency in LLM deployments on mobile devices, as different models exhibit varying responses to prompt characteristics.
Cognitive load can increase energy costs per token by a significant margin, revealing the hidden costs of prompt design in mobile LLM applications.
Large language models (LLMs) are increasingly deployed on mobile devices, making energy efficiency a key deployment constraint, yet the energy impact of prompt design remains underexplored. This paper aims to understand how two prompt properties, cognitive load and phrasing pattern, shape the energy behavior of on-device LLM inference. We conduct a broad empirical study covering prompt properties, datasets, models, and devices, with phase-level profiling that separates prefill and decode energy. We find that cognitive load primarily affects the energy cost per token, while phrasing pattern affects energy largely through token usage. Our energy-quality analysis further shows that prompt design reshapes the attainable frontier differently across models, highlighting the need for model-aware prompt design in energy-efficient on-device LLM inference. Code, datasets, and scripts are available at https://amai-gsu.github.io/PromptProperty/.
