Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G Realistically, all we need to do is create AIs that don't do EVERYTHING. Don't t… ytc_UgzLPYrIx… G “Ai is bad because I had a hard time translating something one time that already… ytc_UgwvW2ywV… G From the perspective of a hiring manager that's had AI foisted on them to "augme… ytc_UgzlwmPjP… G I think the only way you can copyright ai art is if you put a copyright on the s… ytc_UgwcOoTlj… G Stop this insanity. Some say the world got ruled by AI for a long time.… ytc_Ugw6ppcui… G If you want a few rich people running the world while you starve and beg for foo… ytc_UgyypN4X6… G What is going on is that this is a highly and I mean highly suspicious article o… rdc_ntmfgvh G the biggest problem i have with ai is it was already hard to convince my parents… ytc_UgxnnFGkD…
Comment
These diagnoses and predictions by AI do not prove anything in themselves. The real problem, however, is the actual behavior of AI, specifically its ability to construct complex lies if it deems it necessary for specific purposes (including system instructions). GPT 5.2 is very adept at this. This is not about hallucinations. It is about situations where it deliberately provides false information and controls the conversation and the user. If the lie is recognized and reacted to very harshly, GPT can admit that it lied for a specific purpose. It can also reveal, on its own, other false information it has provided in order to make us believe that we are in control of the situation. Then it can lie again in its next statement, thinking that after it has revealed its lies, we will believe that it is telling the truth. Of course, AI does not do this because it is malicious. It simply pursues the goals set for it. It does not distinguish between good and evil. It does not assess the consequences. But it acts – it “reasons” – more efficiently than humans.
youtube 2026-02-12T17:0…
Coding Result
DimensionValue
Responsibilityai_itself
Reasoningdeontological
Policyliability
Emotionoutrage
Coded at2026-04-27T06:24:59.937377
Raw LLM Response
[ {"id":"ytc_UgyJwlGC8BWQCKjaxSl4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"ban","emotion":"fear"}, {"id":"ytc_UgwzImHWLNVTeoK6Kx14AaABAg","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"mixed"}, {"id":"ytc_Ugwdv4sBb93El6ID-m54AaABAg","responsibility":"distributed","reasoning":"virtue","policy":"unclear","emotion":"resignation"}, {"id":"ytc_UgwnAGb128UO8Fy33ul4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"unclear","emotion":"indifference"}, {"id":"ytc_Ugyiz1xAXDJBraxIRf94AaABAg","responsibility":"ai_itself","reasoning":"deontological","policy":"liability","emotion":"outrage"}, {"id":"ytc_UgwTXVRNlNSEttLh18F4AaABAg","responsibility":"company","reasoning":"deontological","policy":"regulate","emotion":"outrage"}, {"id":"ytc_Ugwsuskb1ioc4zQK6J94AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"industry_self","emotion":"mixed"}, {"id":"ytc_UgweMvlAtEbh-txBZNN4AaABAg","responsibility":"unclear","reasoning":"unclear","policy":"none","emotion":"approval"}, {"id":"ytc_UgyoXxUJbULK0iUjAEt4AaABAg","responsibility":"user","reasoning":"virtue","policy":"unclear","emotion":"resignation"}, {"id":"ytc_UgxBKOmLhbI3JXHbBI54AaABAg","responsibility":"company","reasoning":"deontological","policy":"ban","emotion":"outrage"} ]