Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
I just...don't see any evidence of real cognition. It's all well and good to say…
ytc_UgxioM4Nh…
G
As usual, all the normies don’t understand AI at all, and the government getting…
ytc_Ugwozp9vO…
G
how many of us will be facing consequences of climate change and water shortages…
ytr_UgyXEYq34…
G
Liability example: How does AI address potential liability concerns when AI is u…
ytc_Ugyd20kR2…
G
Homework is worthless, I just did it on the bus. It's just to keep you busy.…
ytc_Ugyo7c-8i…
G
You missed the part where all it takes is one psychotic person to misuse AI and …
ytc_UgyaGOPlw…
G
Bro how you going to train a robot how to be slick with them hands 😂…
ytc_UgxT39olR…
G
The guest claims that God is not so good with ethics because evil exists.
Well, …
ytc_Ugzc8sMv8…
Comment
Hallucinations:
I added a saved prompt into Copilot's master list - along the lines of *"<myname> wants factual data unless he explicitly asks for fake or fictional data."* _(would "I" work?)_ And Copilot responded that it understood that the definition of winning would be to give a factual answer or "I don't know" / "I don't have that data" unless I explicitly authorize a fictional response...
I was in the process of adding the same into Gemini, but it wanted to build an exclusive contract about when and where and how... and grew overly complicated... So, I've not implemented the prompt in Gemini yet... But make sure you use the word "winning" as that should trigger the result expectations.
Imagine you ask Gemini to write a short story (fabricate data). Imagine you ask Gemini to summarize a spreadsheet (don't fabricate data). Per Gemini, if it believes the definition of winning is that _a result is expected_ .. then it fabricates a result to win. If you tell it that it cannot fabricate data unless explicitly asked... that maybe stops "hallucinations"... Both Copilot and Gemini said it would. I'm only less than a week in, so no proof or not.
youtube
AI Responsibility
2026-01-13T14:5…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | none |
| Reasoning | unclear |
| Policy | none |
| Emotion | approval |
| Coded at | 2026-04-26T23:09:12.988011 |
Raw LLM Response
[
{"id":"ytc_Ugyt0lvCGjAS8XXiQuV4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"resignation"},
{"id":"ytc_Ugw43gWbkV7n7TkZWa94AaABAg","responsibility":"none","reasoning":"deontological","policy":"none","emotion":"outrage"},
{"id":"ytc_UgwgwZIp5RYHKecEFhB4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_Ugw0EjCY6IigY2r9j3d4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"approval"},
{"id":"ytc_Ugxu0ZC1PynFIZvPTsN4AaABAg","responsibility":"government","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytc_UgyxC0A8pYtC-6NMKZB4AaABAg","responsibility":"government","reasoning":"mixed","policy":"unclear","emotion":"outrage"},
{"id":"ytc_UgzdASnYK7jAowTbgXZ4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_Ugy39yyrzWs6gYtelQh4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"resignation"},
{"id":"ytc_UgwHe25Fk63q4n9wLWl4AaABAg","responsibility":"none","reasoning":"deontological","policy":"ban","emotion":"outrage"},
{"id":"ytc_UgwmPY7qWvaRtBXhwdd4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"approval"}
]