Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Did I hear correctly that US tech guy talking about *needing* to work 60 hours a…
ytc_Ugx1ilvvr…
G
I believe video clip number 2 might be real, the only reason I might say it coul…
ytc_UgzL7qkei…
G
You know how movies that are "based on actual events" have few, if any actual ev…
ytc_UgxBuzSsd…
G
What the fuck is Azmongold's career? Yapping about shit? You can literally have …
ytc_UgwwSI3E0…
G
Oh my god did we go to the same CC. I am full-time online. I literally DREAD my …
rdc_my4zn16
G
You know who has problems with AI? Narcissists. AI is quite simply just the ulti…
ytc_Ugzp0VQ5Q…
G
@charlie7694 That's trivial, because it would be like asking real artists not to…
ytr_UgxTc9Dfz…
G
AI that is observing AI that is observing AI that is observing AI that is observ…
ytc_UgxizrVT4…
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"}
]