Harmlessness Reward Model

该数据集旨在评估生成文本的无害性，它设计了一个奖励模型来完成这一任务。此外，该模型是多目标强化学习算法训练中使用的奖励模型之一。具体任务是对强化学习进行奖励建模。

This dataset aims to evaluate the harmlessness of generated text, and a reward model is constructed for this purpose. In addition, this reward model is among those utilized during the training of multi-objective reinforcement learning algorithms. Its specific task is reward modeling for reinforcement learning.