Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
I can't help but be polite to AI even if I look ridiculous, if I don't, I'll pro…
ytc_Ugznty6Om…
G
Im Currently starting my first year in University and im pursuing a degree in Fi…
ytc_UgxvM4sO1…
G
I dare fucking AI ✨artists✨ that claim that it’s an actual “skill” to put some b…
ytc_Ugxj3-dUD…
G
This answer from ChatGPT)))
Thank you for your comment and we're glad to hear th…
ytr_UgxtdA6Y1…
G
@RC_Edits0 exactly but tell me who in their right mind would do that and even…
ytr_Ugyyo9rOj…
G
doesn't work; this is the reply "I'm really sorry you're going through this. Los…
ytc_UgxNkuRbt…
G
Bro you’re not born with talent, it’s called skill, overtime you become more ski…
ytc_UgywmQMLD…
G
Great video! The two main takeaways:
1) Don't eat bromide.
2) AI represents a…
ytc_UgwmH4s0r…
Comment
That's not the point. The point is that if an AI cannot correctly discern prejudices and understand why they are inherently false, *what other human fallacies might it also accidentally pick up*?
All these dreams people have of the Singularity or even just smaller-scale stuff where AI turns out to be better\smarter than humans necessarily *require* AI to be able to overcome human fallacies. Otherwise, the AI would just end up perpetuating mistakes rather than undoing them. Which could be disastrous, considering how many people seem to actively want machine masters sorting our shit out for us.
Unfortunately, this is the sort of "deeper" understanding of complicated/nonlinear/illogical issues that AI is truly terrible at sorting out, and likely will be for some time to come. It's good that researchers are recognizing that, since at least they can start working on the problem without pretending it's not there.
(But seriously. It can take humans *decades* to learn that some people are just ignorant shitheads AND develop the interpersonal skills to spot and avoid such shitheads. Teaching that to a computer is going to be no small feat.)
reddit
AI Harm Incident
1492371118.0
♥ 3
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | developer |
| Reasoning | consequentialist |
| Policy | none |
| Emotion | fear |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[
{"id":"rdc_dgcci0r","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"rdc_dgc6bg7","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"indifference"},
{"id":"rdc_dgcgdvr","responsibility":"developer","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"rdc_ohuh7jy","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"},
{"id":"rdc_ohueqed","responsibility":"none","reasoning":"consequentialist","policy":"regulate","emotion":"fear"}
]