Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
Yup, it's super hard to analyse speech that is not profane, but is harmful. "Fuck you Motherfucker" is infinitely less harmful to a person than "This is why she left you" but an AI is much better at identifying the former than the latter.
reddit AI Harm Incident 1503155340.0 ♥ 593
Coding Result
DimensionValue
Responsibilitynone
Reasoningconsequentialist
Policynone
Emotionmixed
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_dlullrk","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"indifference"}, {"id":"rdc_dluejq2","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"mixed"}, {"id":"rdc_dlvgbup","responsibility":"developer","reasoning":"mixed","policy":"unclear","emotion":"fear"}, {"id":"rdc_dlun7f7","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"resignation"}, {"id":"rdc_dlucz7i","responsibility":"none","reasoning":"deontological","policy":"none","emotion":"outrage"} ]