Raw LLM — Corpus Dashboard

Look up by comment ID

Random samples — click to inspect

G In a few years many will realize all this hysteria about machine learning was a … ytc_Ugwqchq2Y… G Why would the idiot just stand there and have his hands up like the robot told h… ytc_Ugz3R7jE8… G A fantastic human! 🙏 Thank you for your patience and understanding of the world… ytc_Ugz28idhW… G While people in Waymo are demonstrating how great humans can create great things… ytc_UgyZClvfH… G @mekingtiger9095 The ratio of pretentious AI guys to pretentious "real art" guys… ytr_Ugz32kw6O… G This is why I strongly support hardening ourselves by, every day, speaking to AI… ytc_Ugx6lt36D… G AI will bring down civilization. The harm it is wreaking far outweighs its servi… ytr_UgxZBYqtC… G 28:30 agenda...?? I agrre....re.. What to do with patent level idea initiation… ytc_UgxFI5kcB…

Comment

For those wondering why the model suddenly produced antisemitic output: this is almost certainly a regression caused by optimization, not intent or ideology. In large neural networks, including LLMs, safety behaviors aren’t stored as a separate rule set — they’re distributed across the same parameters that encode everything else. When you fine-tune or otherwise re-optimize the model, you can shift it into a region of the loss landscape where previously learned constraints activate less reliably. That doesn’t excuse the output, but it does explain it. This is a known failure mode in continual learning and model compression, not evidence that the system “became” anything. Treating this as a scare story about AI motives misrepresents what is fundamentally an engineering problem. The correct response is better regression testing, constraint preservation, and robustness — not anthropomorphizing an optimizer.

youtube AI Moral Status 2025-12-15T00:4… ♥ 1

Coding Result

Dimension	Value
Responsibility	developer
Reasoning	consequentialist
Policy	none
Emotion	indifference
Coded at	2026-04-27T06:24:53.388235

Raw LLM Response

[
{"id":"ytc_UgzE2eG0lakVQMLmQtd4AaABAg","responsibility":"ai_itself","reasoning":"mixed","policy":"none","emotion":"approval"},
{"id":"ytc_UgxORgwt8mQqjAxl6bp4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_Ugww3pnixEdYVhaI4-N4AaABAg","responsibility":"developer","reasoning":"deontological","policy":"regulate","emotion":"outrage"},
{"id":"ytc_Ugyi59VSh7djtMh2cqZ4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"},
{"id":"ytc_Ugw1yuRp7mCxI1pCcD14AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_Ugwzx8z2YcMR6qAhw7h4AaABAg","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"outrage"},
{"id":"ytc_Ugw1e5oFOL1yex218WB4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_Ugy-BrjgZBiWFn5wdah4AaABAg","responsibility":"ai_itself","reasoning":"unclear","policy":"none","emotion":"mixed"},
{"id":"ytc_UgwJp4Xy07PkcIaCD-54AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytc_UgwRn0LdjlfiYv_p6Yd4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"regulate","emotion":"outrage"}
]

Raw LLM Responses