Supervised Pretraining Can Learn In-Context Reinforcement Learning | Read Paper on Bytez