Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G It would be factored into the price you pay for service. I think the only conce… rdc_d8a9rxf G I think it’s just people worried about jobs in general. We’re seeing this AI stu… rdc_mr72m52 G 10:43 Google are messing up their own search engine with ai - and now they are t… ytc_Ugz4HMY3l… G Many physical jobs, robots are decades away from doing, not because robots won’t… ytc_UgwtP3ZX6… G This is a fascinating conversation that raises so many questions. One thing that… ytc_UgytsG-LH… G There too good that even there poisoning looks good edit ai is really just a gim… ytc_Ugz8xHAlU… G Okay, long comment from an artist in genuine need of help. I dont use AI, i als… ytc_UgxmhlZ0R… G Why didn't the parents help their son find mental health help? Instead the son t… ytc_Ugzh_3YPl…
Comment
The point about why LLMs hallucinate and don't say "I don't know" at 16:50 is a bit off; it's right, but in a misleading way. The bulk of LLM training is rooted in predicting what word will appear next. If more people wrote "I don't know" online, the LLM would learn to to say "I don't know" more often. The misleading part is that it wouldn't say the phrase because it actually doesn't know, it would say the phrase because it is mimicking what it has seen. The real problem is the fact that LLMs learn mainly from next token prediction, not by interacting with the real world. When a human says I don't know, it is because they have learned from interactions with other humans that saying "I don't know" when they actually don't know is usually a better thing for them to do. As LLMs don't learn in this way, we shouldn't expect them to be able to say when they don't know something.
youtube AI Moral Status 2025-10-30T22:0… ♥ 34
Coding Result
DimensionValue
Responsibilitydeveloper
Reasoningmixed
Policyunclear
Emotionmixed
Coded at2026-04-26T23:09:12.988011
Raw LLM Response
[ {"id":"ytc_UgwiNphKFW9X1-QaJ-14AaABAg","responsibility":"ai_itself","reasoning":"unclear","policy":"unclear","emotion":"mixed"}, {"id":"ytc_UgzRjAa1xY9Z5cAgqhx4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"regulate","emotion":"fear"}, {"id":"ytc_UgzHJqxEZwW92ojEIM54AaABAg","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"indifference"}, {"id":"ytc_UgwKEyRf9Efg1gtDGVN4AaABAg","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"indifference"}, {"id":"ytc_Ugzsz86Dgtuqvi6ELtx4AaABAg","responsibility":"company","reasoning":"deontological","policy":"liability","emotion":"outrage"}, {"id":"ytc_Ugwqp2A-ZgRV4MaerRt4AaABAg","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"approval"}, {"id":"ytc_UgwJPHWUcnvJotZFqnR4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"ban","emotion":"outrage"}, {"id":"ytc_Ugw2h4n1cyMj8mxDYGN4AaABAg","responsibility":"government","reasoning":"consequentialist","policy":"regulate","emotion":"fear"}, {"id":"ytc_UgyAJ2kAfyBWrCvGR6F4AaABAg","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"indifference"}, {"id":"ytc_UgymcRj0Dpo-ThynfKx4AaABAg","responsibility":"developer","reasoning":"mixed","policy":"unclear","emotion":"mixed"} ]