SafetyRating

An assessment of the potential harm of some generated content.

The rating will be restricted to a particular category.

Properties

Indicates whether the content was blocked due to safety concerns.

The category of harm being assessed (for example, Hate speech).

The likelihood of the content causing harm.

A numerical score representing the probability of harm, between 0 and 1.

The severity of the potential harm.

A numerical score representing the severity of harm.