Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
ChatGPT doesn't even predict the next word, it predicts the next token which is a fragment of a word, kind of like a syllable. You can see this if you take a long word and ask it to reverse it. Not sure if GPT4 has fixed this, but in GPT3 it would reverse each token rather than the whole word as expected, creating a mangled result. The point stands though: GPT does not have any understanding of the problem you ask it to solve: it's just using the datasets it was trained on to guess the most appropriate tokens to return using the tokens you fed in.
youtube AI Responsibility 2023-06-11T06:4…
Coding Result
DimensionValue
Responsibilitynone
Reasoningunclear
Policyunclear
Emotionindifference
Coded at2026-04-27T06:26:44.938723
Raw LLM Response
[ {"id":"ytc_Ugw3wSt_iRo4kMpZSF14AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"}, {"id":"ytc_UgwP05cgXoIAKeKmKNZ4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"mixed"}, {"id":"ytc_UgyQpXJVAI1dF-ePbsJ4AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"approval"}, {"id":"ytc_UgyNMwD-JDuUyHH0WjR4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"mixed"}, {"id":"ytc_UgzOMw4cgw6Qv7dnm1l4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"outrage"}, {"id":"ytc_UgyQQvmDk9-pTuke9Cd4AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"approval"}, {"id":"ytc_UgxbkpxCEZfWRtdLHUp4AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"}, {"id":"ytc_UgwtmCYNzOhLvI2pGkd4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"liability","emotion":"indifference"}, {"id":"ytc_UgxExxylhHNjY_1xC3p4AaABAg","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"outrage"}, {"id":"ytc_UgxJ2LzCuYi_iPlB75B4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"mixed"} ]