xqx-chloe1/rollouts_po2_train_2_cleaned_34521

这是一个包含多个环境交互相关特征的数据集，其中包括步骤编号、epsilon值、目标、上一次选择、当前选择、移动速度比例（mpp）、图像路径、智能体位置、朝向、里程表、距离目标距离、目标位置、被拒绝的内容和角色、前沿位置及其ID、到目标和从目标的角度、目标回报、图像列表、选择的内容和角色、提示信息（包括文本和类型）等字段。数据集分为训练集，并提供了默认配置。

This dataset consists of various environmental interaction features, including step number, epsilon value, target, last choice, current choice, movement speed ratio (mpp), image path, agent position, heading, odometer, distance to goal, goal position, rejected content and role, frontier positions and their IDs, angles to and from the goal, target return, list of images, chosen content and role, and prompt information (including text and type). The dataset is split into a training set and provides a default configuration.