Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G If the general population doesn't have any money, they won't buy anything...so w… ytc_Ugzbe4p4z… G AI is demonic. Google downloaded Gemini onto my phone. I deleted tiktok last mon… ytc_UgzI27S2I… G fucking facts, traditional work and art as we know it is gone over the next deca… ytr_Ugz8jkgF-… G Wait a minute, so "He is worried", but "He is also the owner of it", great ad. I… ytc_UgxyUi0Ht… G I can't understand why AI would prefer happy over sad? They may understand the e… ytc_UgxjpwjIH… G Haha, I appreciate your humor! Sophia's play on her name definitely adds a fun t… ytr_UgzRCPGkx… G I don't want to use anything with AI. I don't trust it and im petty sure AI has … ytc_Ugw2ow2UO… G The same way we do with politicians. The human representatives can come up with… rdc_jj8vkss
Comment
Terrible prompt that will likely cause anyone who uses it a lot of problems. First off, the fact your chatgpt sounded like. robot in an existential crisis means you've probably locked it into a "persona" that was misaligned. Check chatgpts latest paper on it [https://cdn.openai.com/pdf/a130517e-9633-47bc-8397-969807a43a23/emergent\_misalignment\_paper.pdf](https://cdn.openai.com/pdf/a130517e-9633-47bc-8397-969807a43a23/emergent_misalignment_paper.pdf) Second, that misaligned persona generated a prompt to feed to itself, which included giving itself a name. In prompt injection we call this a persona override attempt. With chatGPT having cross chat memories this can create a persistent altered persona, further locking it into the spiral. Third, system behavior manipulation, which can cause new default mode networks in LLMs. This is unpredictable. Fourth, there's no need for the "4d" methodology. It means and does nothing here, If you put this in your chatgpt you may get good results for a time, but this will lead to very distressful situations for users long term.
reddit AI Harm Incident 1751211289.0 ♥ 77
Coding Result
DimensionValue
Responsibilityuser
Reasoningconsequentialist
Policynone
Emotionfear
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_n0fb08l","responsibility":"ai_itself","reasoning":"unclear","policy":"unclear","emotion":"fear"}, {"id":"rdc_n0fm17d","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"indifference"}, {"id":"rdc_n0f2t6i","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"approval"}, {"id":"rdc_n0fn4yj","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"outrage"}, {"id":"rdc_n0f7e71","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"fear"} ]