Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
All this AI is bad videos or people saying to themselves that Ai is dangerous, a…
ytc_UgxIxQh8M…
G
So technically the Internet and anything AI is basically channeling a Ouija Boar…
ytc_Ugw6EFLqh…
G
I have full self driving. I was curious if it was worth paying for so i switched…
ytc_UgyqkZj5-…
G
The Internet is virtually unregulated and has become a breeding ground for count…
ytc_UgxtWTQO9…
G
As people increasingly use AI for information, news, entertainment, watch the le…
ytc_UgwzX9KBw…
G
AI cannot be controlled because it is designed to adapt using mathematical compu…
ytc_Ugz5oQZmr…
G
I think the AI is either always agreeing to our inputs because it's programmed o…
ytr_UgxMwNNwd…
G
AI is a sociopath. AI has zero conscience like humans do. It is so stupid to cre…
ytc_UgwuZXcjR…
Comment
The producers may not know how the AI teaches itself, but the programmers know. The fact that they are coy shows that, they were shown how to improve at chess, and simply applied lots of computation. So for chess we now know, the software ranks all the possible moves using a points system. It assigns VARYING weights to different factors such as centrality to the board, values of the pieces, opening up the position, etc, and runs millions of games. It can see from the data that it plays better chess when it assigns X to this component of the move and Y to this othe rcomponent. So it is honing in on how to properly evaluate each move. Grand masters used to argue the pros and cons of various possible moves and they had their own philosophies. Well the computer puts hard numbers to all this and learns to beat a grand master by creating a philosophy backed up and informed by winning percentages, and playing a billion times more games than a grandmaster could ever play.
The same techniques were used to train the chatbots. The feedback is more subjective and nebulous but you train a machine in the same old ways, you show it how to improve in a few examples, and then it fills in the millions of blanks by itself.
youtube
2025-07-06T01:3…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | developer |
| Reasoning | consequentialist |
| Policy | unclear |
| Emotion | indifference |
| Coded at | 2026-04-27T06:24:59.937377 |
Raw LLM Response
[{"id":"ytc_UgxHx-HIU4MS7LvHZZ54AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"approval"},
{"id":"ytc_UgwF-2tUx8mRp-AUukt4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_UgykdKH261UZSDKRrup4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_UgyfAjDvskg7UtySrjV4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"unclear","emotion":"indifference"},
{"id":"ytc_UgxwTkJoi_Deyv_f_O54AaABAg","responsibility":"ai_itself","reasoning":"unclear","policy":"none","emotion":"mixed"},
{"id":"ytc_UgxZIdgf9JDniqQEKSt4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_UgzAhrwvJWpLMU53HQx4AaABAg","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"approval"},
{"id":"ytc_UgxEfYQrsF1Uf4kEvah4AaABAg","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"mixed"},
{"id":"ytc_UgzvI-jxfbCo6jKz_YZ4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"ytc_UgznRz2DOBARvqUPU6B4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"ban","emotion":"outrage"}]