@misc{indiciaef0328339e001,
  title = {Bridging the Training-Inference Gap in LLMs by Leveraging Self-Generated Tokens},
  author = {Zhepeng Cen and Yao Liu and Siliang Zeng and Pratik Chaudhari and Huzefa Rangwala and George Karypis and Rasool Fakoor},
  year = {2025},
  url = {https://arxiv.org/abs/2410.14655},
  note = {Source identifier: 2410.14655}
}