This is a PureJAX version of our NeurIPS 2023 paper "Structured State Space Models for In-Context Reinforcement Learning". We evaluate and modify S4-like models for reinforcement learning. Furthermore ...