Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G Close to 50% of of total American companies are not hiring new workers this 2025… ytc_UgwgwVBjA… G We're glad you found the interaction intriguing! If you're curious to delve deep… ytr_UgwzT7xA9… G I'm one of those people, i used to be able to draw (mediocre) but for the past f… ytr_Ugz679GRD… G Great questions! In the video, Sophia emphasizes her connection to wisdom and he… ytr_UgzaS9CSr… G I feel if the status quo were without AI, humanity actually is regressing and ov… ytr_UgypAiNi4… G I’m seeing the same comment over and over again saying it’s a ‘slave shortage ‘.… ytc_UgwTCMIzQ… G Ai works faster and then human corrects ots mistakes and then humans train ai to… ytr_UgywwSr9P… G AI caused reorg on my team like 10 times over the past 3 years or so. Couldn't s… ytc_UgxCII6Ni…
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"} ]