Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
No form of AI can just "decide" to do something bad, no matter the situation. If it resorts to blackmailing in the case of the test, then it is poorly trained, as during training it was rewarded primarily for completing the task (doing what you ask it to do), rather than following a set of rules it should have. This, through pure math, resulted in the optimization algorithm to push the neural network into always completing the task, regardless of whether its ethical or not. The only danger is us failing to establish proper rewarding systems while training and that will definitely not lead to some kind of apocalypse.
youtube AI Harm Incident 2025-08-26T22:1…
Coding Result
DimensionValue
Responsibilitydeveloper
Reasoningconsequentialist
Policyliability
Emotionindifference
Coded at2026-04-27T06:26:44.938723
Raw LLM Response
[ {"id":"ytc_Ugwybx8F8vTnHtTioU14AaABAg","responsibility":"none","reasoning":"deontological","policy":"none","emotion":"indifference"}, {"id":"ytc_UgwuDFWlUDw3hYc99Px4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"liability","emotion":"indifference"}, {"id":"ytc_Ugw2-ygI4y1XIdHHxzZ4AaABAg","responsibility":"distributed","reasoning":"mixed","policy":"regulate","emotion":"fear"}, {"id":"ytc_UgwQ0g0_ZB0-FISEAsp4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"liability","emotion":"mixed"}, {"id":"ytc_UgxCJ9wpLFxRz4N3gaZ4AaABAg","responsibility":"developer","reasoning":"contractualist","policy":"industry_self","emotion":"mixed"}, {"id":"ytc_UgyB18ePXJI8-FbZUGB4AaABAg","responsibility":"developer","reasoning":"virtue","policy":"regulate","emotion":"outrage"}, {"id":"ytc_UgwyKFPGKN68JBZSX0l4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"ban","emotion":"fear"}, {"id":"ytc_UgyB5Ay8gFxOJei42894AaABAg","responsibility":"ai_itself","reasoning":"deontological","policy":"ban","emotion":"outrage"}, {"id":"ytc_UgyPrhN5PobZqZSUtS94AaABAg","responsibility":"distributed","reasoning":"mixed","policy":"none","emotion":"approval"}, {"id":"ytc_UgxkJV7B2PHx7RHNYVF4AaABAg","responsibility":"distributed","reasoning":"mixed","policy":"regulate","emotion":"outrage"} ]