Training Power Grid Standard LLMs via Token-Adaptive Continual Pretraining and Orthogonal Subspace On-Policy Self-Distillation
Zhiyuan Hu; Chengwei Peng; Zhichen Peng; Zhuojun Cai; Zhichao Zhang; Zhao Sun; Jianming Hu; Shouyu Liang
- Published
- Publisher
- MDPI AG
- Journal / collection
- Mathematics
- Research area
- Energy & infrastructure
- Source type
- Journal article
- DOI
- 10.3390/math14193479
The original publisher provides the full text and its access options.
Cite this source
Zhiyuan Hu; Chengwei Peng; Zhichen Peng; Zhuojun Cai; Zhichao Zhang; Zhao Sun; Jianming Hu; Shouyu Liang. Training Power Grid Standard LLMs via Token-Adaptive Continual Pretraining and Orthogonal Subspace On-Policy Self-Distillation. Mathematics. 2026-09-24. https://doi.org/10.3390/math14193479