Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
True and it's still painful to remove background from a pic to make a perfect pa…
rdc_n7yolgh
G
I would be lying if I said I didn't see the value in this. I could totally see m…
ytc_UgxyxI3sZ…
G
Ai images is like putting self driving training wheels on a bicycle while art is…
ytc_UgzkfWceG…
G
Humans are "clay". AI is "iron"....the 2 will not mix.
Where oh where have we h…
ytc_Ugyctc__3…
G
@paulster185 Yea that's the problem. AI can be intelligent and extremely useful …
ytr_Ugxkjh_uS…
G
The car manufacturer gets sued because it's their creation. It was their decisio…
ytc_UgjcFgyqV…
G
Let AI take over Congress and start at the top of the White House, then work …
ytc_Ugx2J_0pj…
G
The big problem I have is that the official script is a summary of every answer …
ytc_UgxDUFUTT…
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"}
]