Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
My girlfriend fights exactly like you do, it's exhausting. Poor chatgpt 😂
Jk, g…
ytc_Ugylt_D3r…
G
Thank you for watching our video! If you're interested in more interactive sessi…
ytr_Ugyf7O6ZA…
G
They wont need us soon. The billionaires are already developing autonomous AI mu…
ytr_UgxAzUqw7…
G
We appreciate your feedback and understand your concerns. If you're interested i…
ytr_UgxeTvJt9…
G
Let’s be frank. Ai is the future. Let me be the first to bow down to our AI over…
ytc_UgxXlA02l…
G
Something to understand about AI. It is not simply a "tool" to make your life ea…
ytc_UgwhvvZiN…
G
So what is the rest of the story about the threats from Claude 4? ...? Did they …
ytc_Ugw4tnbRx…
G
The thing about AI is that it learns from being plugged into the internet so ess…
ytc_UgymnWYml…
Comment
From the article and the images showing the process flow, it shows up as a legitimate Copilot link. So some people parsing the URL might think "oh that's Microsoft, nothing on their service should harm me" and that's where the flaw is. They engineered the prompt to ignore existing Copilot safeguards against this.
reddit
Cross-Cultural
1768411257.0
♥ 6
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | company |
| Reasoning | deontological |
| Policy | regulate |
| Emotion | fear |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[
{"id":"rdc_nzkxf2c","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"unclear"},
{"id":"rdc_nzkhrxr","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"unclear"},
{"id":"rdc_nzklucd","responsibility":"company","reasoning":"deontological","policy":"regulate","emotion":"fear"},
{"id":"rdc_nzkeae5","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"rdc_nzkd8ly","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"outrage"}
]