Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
This is great news for alignment, monitoring the model states can be used to prevent misalignment tuning things like self preservation, the AI acts misaligned not because it wants to live it don't have real feelings is acting according to what it learned from the training data (human text with human emotions)
youtube AI Moral Status 2026-04-21T04:4…
Coding Result
DimensionValue
Responsibilityai_itself
Reasoningconsequentialist
Policyregulate
Emotionapproval
Coded at2026-04-26T23:09:12.988011
Raw LLM Response
[ {"id":"ytc_UgyZm78H4Oy_jISHnLt4AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"}, {"id":"ytc_UgzZMCHZLE5IrNPZ0tp4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"regulate","emotion":"approval"}, {"id":"ytc_UgwCDQJUoN3vsyoOAgt4AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"mixed"}, {"id":"ytc_UgwksPggIwOO1lmTW3l4AaABAg","responsibility":"government","reasoning":"mixed","policy":"liability","emotion":"outrage"}, {"id":"ytc_Ugygs_-znj5gWhgx9It4AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"approval"}, {"id":"ytc_UgxHyQ3_TPTRnEwtCp94AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"approval"}, {"id":"ytc_Ugz7GXW57ovo3J172MN4AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"resignation"}, {"id":"ytc_Ugw37AsNcdHpsFrPdLx4AaABAg","responsibility":"user","reasoning":"contractualist","policy":"none","emotion":"approval"}, {"id":"ytc_UgwmPhoehPgw6KTptB94AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"regulate","emotion":"approval"}, {"id":"ytc_Ugy0bKxsqJJQl9z5WNF4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"ban","emotion":"fear"} ]