Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
The words "without being asked" are really doing work in that headline. it implies it was generating these out of complete nowhere, like when the previous times with Grok where it spouted racist stuff unprompted. But this is literally what the author asked for, indirectly. This is the kind of promptings people do when they want nude in Midjourney but trying to bypass the filter.
reddit AI Harm Incident 1754471691.0 ♥ 16
Coding Result
DimensionValue
Responsibilityuser
Reasoningconsequentialist
Policyindustry_self
Emotionmixed
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_n7527vd","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"}, {"id":"rdc_n74e7em","responsibility":"distributed","reasoning":"consequentialist","policy":"liability","emotion":"outrage"}, {"id":"rdc_n76hwfq","responsibility":"user","reasoning":"consequentialist","policy":"industry_self","emotion":"mixed"}, {"id":"rdc_n74lqxd","responsibility":"ai_itself","reasoning":"deontological","policy":"regulate","emotion":"outrage"}, {"id":"rdc_n778q3d","responsibility":"user","reasoning":"consequentialist","policy":"industry_self","emotion":"mixed"} ]