GitBag/llama3-ultrafeedback-reasoning-iter_2-1732268914-armo

该数据集包含多个轨迹的解决方案和思考过程，每个解决方案都有一个对应的奖励值。数据集的字段包括提示（prompt）、增强提示（augmented_prompt）、多个轨迹的解决方案和思考过程（traj_X_solution_Y和traj_X_thought_Y），以及每个解决方案的奖励（traj_X_solution_Y_reward）。数据集包含一个训练集，共有60814个样本，文件大小为1505264794字节。

This dataset contains solutions and thought processes for multiple trajectories, with each solution having a corresponding reward value. The dataset fields include prompt, augmented_prompt, multiple trajectory solutions and thought processes (traj_X_solution_Y and traj_X_thought_Y), and rewards for each solution (traj_X_solution_Y_reward). The dataset includes a training set with 60,814 samples and a file size of 1,505,264,794 bytes.