Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
T yt Jerez a difference between being inspired by and down right copying. Since …
ytc_UgwO5TpT-…
G
they had Beemo, and the butter robot from Rick and Morty as well.
gotta love all…
ytr_UggJWv_0c…
G
The globalists want to eliminate six billion people from this planet by 2030. Th…
ytc_UgyvnwY32…
G
I just listened to Diary of a CEO podcast with a pioneer of AI, Geoffrey Hinton.…
rdc_n0i8t3c
G
AI CEO: Some of you may die but that's a risk my bank account is willing to tak…
ytc_UgxlEeLtC…
G
If time has finally come when we are supplanted by our AI offspring, why resist?…
ytc_Ugzue5KXc…
G
ITS MATH AT SCALE APPLIED TO OBJECTIVES.
its accuracy - its about what humans p…
ytc_Ugx_cVaBA…
G
Is he telling me that we can’t put a navigator system in a robot and drop them o…
ytc_Ugx14GL-G…
Comment
> Why have AI's been trained on data that includes some of the most imaginative ways for an AI to destroy and supercede humanity ever conceived? Wouldn't giving it extremely dangerous "ideas" about what it is and could be capable of be the FIRST thing they should be working to remove from the training data and reverse course on? You know... the data it's using?
The answer to this is pretty straightforward. Current systems are extremely unlikely to seek or be able to effectively carry out an attack, while future systems are extremely likely to encounter this information, or derive it on their own. Therefore there's little to no risk in teaching current systems about these things, and doing so may help mitigate the risk in future systems as we can use current systems in deconfusion research to help us align future systems.
reddit
AI Moral Status
1685650521.0
♥ 6
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | developer |
| Reasoning | consequentialist |
| Policy | regulate |
| Emotion | fear |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[
{"id":"rdc_jmiu1k6","responsibility":"developer","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"rdc_jmiyavu","responsibility":"distributed","reasoning":"mixed","policy":"none","emotion":"resignation"},
{"id":"rdc_jmfrnw7","responsibility":"media","reasoning":"deontological","policy":"none","emotion":"indifference"},
{"id":"rdc_jmfyo7p","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"mixed"},
{"id":"rdc_jmi5ky3","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"outrage"}
]