Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Thank you for sharing your thoughts! It's true that many people value human conn…
ytr_UgwQdkr2f…
G
He doesn’t care about the consequences, he wants enough regulation that it’s too…
ytc_Ugxgs659S…
G
art is about the process and thought behind it. ai doesnt have a brain i therefo…
ytc_UgxjQVtsc…
G
it will make humans worthless. Yes, money matters. But its about the satisfactio…
ytc_Ugw8409Zo…
G
@whilpin I think defining the “point of art” as an “overall idea or fee…
ytr_Ugx8fCyMf…
G
Man, this feels like the laziest hit piece I have ever watched. A few points:
1.…
ytc_Ugwm5s3Lm…
G
I love these AI ones and about business but I wish you guys would interview more…
ytc_Ugwjn5n0Z…
G
I am a pianist and give private lessons. No way AI can give music lessons to ind…
ytc_Ugwb_6Xzl…
Comment
Re: How the model prioritizes, I think the word they're looking for is heuristics (common term in literature). In other words, defining the rule that says "Yes you trained well!" The biggest problem with training is developing a sensible heuristic. Model parameters and architecture affect training efficiency and fit, but models basically gamify whatever the creator says gives it highest score. If lying is the optimal way to score better, then it will do that. So if our scoring system says we want better similarity between the generated text and training text, then it will do that by whatever means gets the score the highest. Mathematically modeling correctness and humanness of text is not an easy thing.
(Disclaimer, I am a PhD student that uses machine learning but not an LLM researcher)
youtube
AI Moral Status
2025-11-12T19:0…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | developer |
| Reasoning | consequentialist |
| Policy | none |
| Emotion | indifference |
| Coded at | 2026-04-26T23:09:12.988011 |
Raw LLM Response
[
{"id":"ytc_UgylT8svfl2oMUW4U-F4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytc_UgyLfTtZm68taB7U9cp4AaABAg","responsibility":"company","reasoning":"virtue","policy":"none","emotion":"outrage"},
{"id":"ytc_UgxN-eoii1kT-akvwbF4AaABAg","responsibility":"developer","reasoning":"mixed","policy":"none","emotion":"approval"},
{"id":"ytc_Ugyxl84xUl_8ihgo6Oh4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytc_UgyI7zZSTLvif6F1Eex4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytc_Ugy_r8BM6oN8TggtiKZ4AaABAg","responsibility":"ai_itself","reasoning":"mixed","policy":"none","emotion":"approval"},
{"id":"ytc_Ugw-fujppI_piFchIax4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_UgwpWiqiuGSZK0WrC594AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"},
{"id":"ytc_UgzoqF7ccItFYIIxCMl4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},
{"id":"ytc_Ugxj-Qku7l1HM-CHoU94AaABAg","responsibility":"distributed","reasoning":"consequentialist","policy":"regulate","emotion":"fear"}
]