@article{pelmpowerefficientondevicellminferencewi,
  title = {PELM: Power Efficient On-Device LLM Inference with Speculative Decoding and Dynamic Voltage Frequency Scaling},
  author = {Weisi Yang and Stephen Xia},
  year = {2026},
  eprint = {2609.09662},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.09662},
}