Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G Whoa that is just crazy but sadly I think this guy would probably do it anyway i… ytc_UgxJomGbt… G I only use ai art for the haha funny share with friends saying "lol look at this… ytc_Ugwo07u9J… G I use AI art for my classes. The copyright laws around educational use are more … ytc_Ugx67B6mN… G Both humans and ai have racism problems But only humans can make it a funn… ytc_Ugxuy2-YI… G It's one thing to mistake ChatGPT for a search engine. But jeez, the chain the i… ytc_Ugz9oHtJx… G Y'all got chats you don't want anybody to see, while I'm here just using charact… ytc_UgxD6yFwA… G Human greed will eventually loose and A.I will win, inevitable, laws will not he… ytc_Ugy47YrQA… G I'm a PM and AI has made my life so much easier, work smart, not hard people.… ytc_Ugz3udchk…
Comment
I doubt that much of this has made it into the actual training data of the models. Since the article is mentioning testing "chat bots", most of their results are probably from the models doing a web search and paraphrasing, with the providers not doing much to prevent misinformation. Think of that what you will.
reddit AI Harm Incident 1743171346.0 ♥ 17
Coding Result
DimensionValue
Responsibilitycompany
Reasoningconsequentialist
Policyliability
Emotionmixed
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_mk6dlbs","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"}, {"id":"rdc_mk6rnxp","responsibility":"company","reasoning":"consequentialist","policy":"liability","emotion":"mixed"}, {"id":"rdc_mk6x9fb","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"indifference"}, {"id":"rdc_mkba3m2","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"outrage"}, {"id":"rdc_mk7mhoh","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"mixed"} ]