##article.return## Reversing Large Language Models for Efficient Training and Fine-Tuning Download Download PDF