Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Bro there was this one this one stan that liked deku from my hero academia on re…
ytc_UgxDPjPRh…
G
Bro ai art is the same as actual art most ppl only look at the results like why …
ytc_UgxtfQVIe…
G
“Godfather of ai” is a ridiculous name. It gives him a false level of authority …
ytc_UgzvCydqA…
G
I'd say the Mavis one isn't AI simply because the shot moves and she actually do…
ytc_UgxMY1I8h…
G
You’re misrepresenting the findings. By From the study’s FAQ:
Is it safe to say…
ytc_Ugzn8E64i…
G
Bearing in mind that every ladder set proposes a different obstacle exterior con…
ytc_UgzFfxzth…
G
Is this a video with real Optimus units and a real Cybertruck or is it an AI gen…
ytc_UgxGImbEB…
G
Unless humans ban together and stop using any companies that use AI.
I can see …
ytc_UgxXnm2gx…
Comment
It just tells you that while we are getting very advanced with AI, they still can’t understand a deeper context why something is the way it is.
The thing is with Gemini, google put a “safeguard”, but it just gave them an unexpected outcome. That being said, something like this shouldn’t have slipped QA.
Put it simply, being racist towards white has a more “acceptable” outcome compared to when it is racist towards, black, poc or etc which can even lead to boycotts or that kind of shenanigans.
reddit
AI Harm Incident
1708881416.0
♥ 79
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | company |
| Reasoning | deontological |
| Policy | liability |
| Emotion | outrage |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[
{"id":"rdc_ks2nwb7","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"rdc_ks3ai0r","responsibility":"company","reasoning":"deontological","policy":"liability","emotion":"outrage"},
{"id":"rdc_ks2kfws","responsibility":"ai_itself","reasoning":"mixed","policy":"none","emotion":"approval"},
{"id":"rdc_ks2y5lb","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"outrage"},
{"id":"rdc_ks4es5f","responsibility":"ai_itself","reasoning":"deontological","policy":"none","emotion":"fear"}
]