range of perference scores

[ctxya1207]

Could you please tell me what is the range of preference scores generated by ImageReward?

[xujz18]

The reward is normalized to have a mean of 0 and a standard deviation of 1. When testing stable diffusion v1.4 on our metric set, the scope of the reward is observed to have 62.4% of [-1,1] and 98.2% of [-2,2].
You can refer to: Issue 5, Issue 7, Issue 22, Issue 40, and Issue 73.