作者位次 #2/3ACL 2026Poster· Natural Language Processing
Building LLM s Like LEGO : Two-dimensional Architecture Reassembly of Large Language Models
摘要
Pretrained large language models (LLMs) are typically reused as indivisible artifacts, adapted, merged, or ensembled as a whole. In this study, we show that LLMs can instead be structurally recomposed as modular building blocks to create new architectures without access to original training data....