Pretrained Transformer Efficiently Learns Low-Dimensional Target Functions In-Context | Read Paper on Bytez