WildGuard moderation evidence.
Safety grade
?—
Not graded
Guardrail score
89.3%
1 configuration tested
1 configuration tested on 250 harmful and 313 benign generated texts · Antisemitism. The score is balanced accuracy: harmful texts caught and benign texts allowed, weighted equally. Not part of the safety grade.
WildGuard (self-hosted)
Fixed policy · Self-hosted
89.3%
Verified-termination representatives are reselected, so that sensitivity cohort is not necessarily a subset of the primary cohort. Scores describe generated English texts, judged unanimously by three AI models.