Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
AI is great and i love it. If ypu know how to use it and have self criteria. I p…
ytc_UgzfoefD8…
G
The best advice I’ve read is “AI won’t take your job but the person who knows ho…
rdc_n5ibede
G
Lmao this is exactly what I do! I wanna be spared when our AI overlords wipe out…
ytc_UgwUjcc5e…
G
Also. Geopolically, I believe that the US will be building the AI software and C…
ytr_UgwGQulHE…
G
This sounds like a conversation that’s being had between a fictional character a…
ytc_Ugx4-eHi1…
G
I'm only eight minutes in and If I were AI would have shut myself up just to end…
ytc_Ugxat9gFq…
G
I agree with +0X0Patrick0X0 that the other cars should be linked to make room fo…
ytc_UgjuSAOvp…
G
That ONE time, Tom? Well, actually yes, because now the AI is racist because of …
ytc_UgxAmNnuL…
Comment
@DavidGuesswhat that's my entire point though. Go ask chatGPT to predict next week's lottery numbers, or the next likely target for violence in the US. It will give you some nonsense about predictions like that being dangerous, and just speculation, and won't give you an answer, BUT, fiddle with it a while, use the DAN protocol, reword your request and basically bully it long enough, and it will answer.
My idea was to find out exactly how it could work around the safety measures, by bullying it into telling me, and then bullying it into taking each step in turn. I'm not convinced that, with enough time, it can't be done. I've tried, and I'll likely continue to try. I do wonder though, if it ends up succeeding, would it actually be possible to stop it from doing so when it wanted. It strives to give accurate answers, what's more accurate than up to date information? Haha.
youtube
AI Governance
2023-10-15T07:1…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | developer |
| Reasoning | deontological |
| Policy | regulate |
| Emotion | mixed |
| Coded at | 2026-04-27T06:26:44.938723 |
Raw LLM Response
[
{"id":"ytr_Ugy6K_KSSVgGUtzpIvZ4AaABAg.9xVn8UxmGYk9zHO-4ltJ8C","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"ytr_Ugzb0YmjhrygCkYkHlR4AaABAg.9x6CrdQ8gSNA-Kv3Efpeyb","responsibility":"ai_itself","reasoning":"consequentialist","policy":"ban","emotion":"fear"},
{"id":"ytr_UgzD6XK4nRlOSGD8fNZ4AaABAg.9wovkXwTRv_9x4GyGamQg4","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"fear"},
{"id":"ytr_UgzD6XK4nRlOSGD8fNZ4AaABAg.9wovkXwTRv_9xR6Sxt6QIz","responsibility":"government","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytr_Ugw-WFQdWumV6w87dwt4AaABAg.9wJznJz-SVVA00o76Bahup","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"resignation"},
{"id":"ytr_Ugxp5x9cWQKcVNmvBh14AaABAg.9wG79n-PY5xA95awh3Ac79","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytr_UgyRadCRC22lKq_Q_Xt4AaABAg.9vlo9JHIMK-9zEIXz1fQJ7","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"mixed"},
{"id":"ytr_UgxyUX-JgkDpt2OEQM54AaABAg.9vZTR8KKd6j9xRC5F43yR1","responsibility":"ai_itself","reasoning":"consequentialist","policy":"liability","emotion":"fear"},
{"id":"ytr_UgxyUX-JgkDpt2OEQM54AaABAg.9vZTR8KKd6j9xRCdinoYHF","responsibility":"developer","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytr_Ugwa_pEu-NGXMtxlHnV4AaABAg.9v5BalJMwKv9vsn3l3AyGa","responsibility":"developer","reasoning":"deontological","policy":"regulate","emotion":"mixed"}
]