1 paper · 1 filter
Keyao Zhang, Yiquan Chen, Zhuo Hu +3
The accuracy of large language models (LLMs) improves with increasing model size, but increasing model complexity also poses significant challenges to training stability. Periodic…