Intern, Google Deepmind
1 paper at NeurIPS 2025
We investigate an efficient and effective context-window scheduling method for language model pretraining.