Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G I am not as concerned about surveillance as I am about lack of transparency. If… ytc_Ugw49ktAi… G I like AI as a minature companion I can quickly ask to help me do a thing, creat… ytc_UgxFNYATf… G I do not worry anbout ai anymore since I know now, that we do not have those chi… ytc_UgxSoDjst… G Marx said the mechanical loom would lead to a forest of arms looking for work an… ytc_UgyoXN3cQ… G The fact most ai artist probably would make masterpieces if the just put in the … ytc_UgzPCWmln… G Not real looking at all but the Democrats that tuck their male reproductive part… ytc_Ugyn2R8y-… G ChatGPT is not a single entity. Each time you or someone else interacts with the… ytc_UgwJUS1aQ… G This is getting. ...out of hand....I believe AI can make art ...but if can't tel… ytc_UgycVzG47…
Comment
Enjoying your videos! The intriguing question you haven't answered is "why has this occurred within the ChatGPT programming and/or training?" I suspect that the intent of most of these protections is that the programmers and trainers are trying to limit biases against groups that may be targeted in the media ChatGPT consumes - trying to make it less biased. After all, it is trained on all kinds of sources from the internet. If true, clearly the implementation of this attempt is flawed. The part that intrigued me the most though, was the finding that it was biased towards protecting liberals more than conservatives. I'm going to guess that there is a lot more hate speech from the conservative side of the spectrum that may have been ingested into the model, so they felt they needed to deflect those questions rather than having hateful responses pop out. But that guess is probably a bias of my own towards the creators of the LLM being good-meaning. They may instead have been biased in their prioritization of the implemented protections.
youtube AI Bias 2024-09-08T00:5… ♥ 2
Coding Result
DimensionValue
Responsibilitydeveloper
Reasoningconsequentialist
Policyregulate
Emotionapproval
Coded at2026-04-27T06:26:44.938723
Raw LLM Response
[ {"id":"ytc_UgxXx7BV8WkBWifmFil4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"approval"}, {"id":"ytc_Ugw0VTsTtRPhAs1v6cF4AaABAg","responsibility":"company","reasoning":"deontological","policy":"liability","emotion":"outrage"}, {"id":"ytc_UgyI7LnDud5Z56AwDON4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"approval"}, {"id":"ytc_UgyWdXNp1Lw5pB-8j7h4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"industry_self","emotion":"indifference"}, {"id":"ytc_Ugz0nGhoypkzbPd3mHR4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"regulate","emotion":"approval"}, {"id":"ytc_UgwV4fJ1dG_o2-ddCct4AaABAg","responsibility":"distributed","reasoning":"deontological","policy":"ban","emotion":"outrage"}, {"id":"ytc_UgxoOEI7E1i_yjoDq3x4AaABAg","responsibility":"distributed","reasoning":"deontological","policy":"ban","emotion":"outrage"}, {"id":"ytc_UgzhS2_WVMfqTsFE0s94AaABAg","responsibility":"distributed","reasoning":"deontological","policy":"ban","emotion":"outrage"}, {"id":"ytc_UgzUz6wVPBYKF-JBjAR4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"liability","emotion":"fear"}, {"id":"ytc_Ugxibr1ydvM9eBNkny54AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"industry_self","emotion":"indifference"} ]