LoGAH: Predicting 774-Million-Parameter Transformers using Graph HyperNetworks with 1/100 Parameters
arXiv
May 31, 2024
Stacking Your Transformers: A Closer Look at Model Growth for Efficient LLM Pre-Training
NeurIPS 2024 Spotlight
May 30, 2024
AI Alignment: A Comprehensive Survey
arXiv
Oct 30, 2023
Interactive Natural Language Processing
arXiv
May 22, 2023