English

Training LLMs Beyond Next Token Prediction -- Filling the Mutual Information Gap

Computation and Language 2025-11-04 v1 Artificial Intelligence

Abstract

Optimizing training performance in large language models (LLMs) remains an essential challenge, particularly in improving model performance while maintaining computational costs. This work challenges the conventional approach of training LLMs using next-token prediction (NTP), arguing that by predicting information-rich tokens during training, there is a more effective way to train LLMs. We investigate the impact of the proposed solution in three kinds of tasks for LLMs: arithmetic, multi-label classification of text, and natural-language generation. This work offers a principled approach to optimizing LLM training, advancing both model performance and theoretical understanding of the target-token selection strategies.

Keywords

Cite

@article{arxiv.2511.00198,
  title  = {Training LLMs Beyond Next Token Prediction -- Filling the Mutual Information Gap},
  author = {Chun-Hao Yang and Bo-Han Feng and Tzu-Yuan Lai and Yan Yu Chen and Yin-Kai Dean Huang and Shou-De Lin},
  journal= {arXiv preprint arXiv:2511.00198},
  year   = {2025}
}