Raw LLM — Corpus Dashboard

Look up by comment ID

Random samples — click to inspect

G Ai needs to go now someone created this shit someone can get rid of it. It needs… ytc_UgzALScK3… G Why are they creating these? He said that humans need to be smarter. Humans have… ytc_Ugx5XKZ0Z… G This reminds me of videos on YouTube of talking green parrots. They look like th… ytc_Ugz-aSb8V… G I think it deserves to not be shut down... who wants that? If it's modeled after… ytc_UgwR9jxVB… G I think AI is a bad decision in the first place. We've all seen what happens in … ytc_UgiltTSEW… G Not even the em-dashes, he overuses the hell out of superlatives and two-part se… ytr_UgzOTC2_C… G This guy seems like the man tbh. Recognizing that the AI catch tools arent real … rdc_nu11w30 G I agree with part of this especially in the future but for now Ai art has imperf… ytr_UgyoHvErX…

Comment

I work on AI and physics research. While we have made so much advancements, the so called "emergent" properties of AI, such as coding, is little exaggerated by media. Researchers were surprised at first because smaller models couldn't even write a simple functional snippet but as they increased the parameters and training data, this ability appeared to write functional code suddenly popped up. Although it's a bit controversial whether they are truly emergent or not among researchers. The training data still has code, so it's not like they magically gained the ability with size out of nowhere. What really improved drastically was their pattern recognition skill, that's what makes them write flawless code even without actually executing it. I should point out that today's premier models still lack spatial and temporal awareness compared to humans. Even multimodal LLMs like Gemini Pro which is the best at vision struggle with simple geometry problems. The reason is that their training data lack diversity of such simple problems. True intelligence is when the skills/experience/knowledge you learn to tackle a specific set of problems can be mapped to solve a completely different set of problems. This also requires recognizing patterns, but I believe AI will reach there at some point.

youtube AI Governance 2026-03-17T02:1… ♥ 18

Coding Result

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

Raw LLM Response

[
{"id":"ytc_Ugz-QyXSsGQeMWFQ1VV4AaABAg","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"approval"},
{"id":"ytc_Ugy15kxUw5OCMZEMrQB4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"indifference"},
{"id":"ytc_Ugxrejwji932AoTruw14AaABAg","responsibility":"none","reasoning":"deontological","policy":"none","emotion":"outrage"},
{"id":"ytc_UgwJfLLPePKZVeDHAJ14AaABAg","responsibility":"developer","reasoning":"virtue","policy":"none","emotion":"outrage"},
{"id":"ytc_Ugxu4b0ilvG05zZZLmN4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"resignation"},
{"id":"ytc_Ugx5A_nVzUJcyFM7HQh4AaABAg","responsibility":"ai_itself","reasoning":"mixed","policy":"none","emotion":"mixed"},
{"id":"ytc_UgyyPvD8C9P2L3zKadZ4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytc_UgxBVjyRX7qRk2Hq0t14AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"outrage"},
{"id":"ytc_UgzfMM8gnINCIFNDEPt4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytc_Ugy6ob7PvE4UnWyaWWV4AaABAg","responsibility":"distributed","reasoning":"contractualist","policy":"regulate","emotion":"fear"}
]

Raw LLM Responses