Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Without jobs there will be no money to use automated uber or other services
……
ytc_UgxTnsusf…
G
what du you mean ai or robots is far way more dangerous than nukes it's like not…
ytc_UgzsV9y6M…
G
I do get what you’re saying. But it was a much different time, back then.…
rdc_ess8zic
G
I could, theoretically imagine a point where "AI" (actual AI, not what corporate…
ytc_UgyKuPqqp…
G
AI should be way more restricted, and we have to be very careful bc one day the …
ytc_UgyHm9HfS…
G
Nintendo has come out and said they have no intention of ever utilizing generati…
ytc_Ugz7e5cLy…
G
If you ask the AI "√9 = ?" and "√9 = 3" is the answer, you don't say "See? I cal…
ytc_Ugz3LPDHW…
G
AI and mexicans are the same stuff:
everybody finds it awesome.
But you can't …
ytc_UgzMsWxov…
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"}
]