1 paper
Yingjie He, Zhaolu Kang, Kehan Jiang +17
Large language models (LLMs) excel at semantic understanding, yet their ability to reconstruct internal structure from scrambled inputs remains underexplored. Sentence-level restor…