Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G And not just the leadership. Every idiot who voted for the pedophile POS should… rdc_oi0n84x G Elon said he's building a Internet in Space. Should he create an ICANN TLD DNS … ytc_UgxIUWRBN… G Saying 'LLMs are just fancy autocomplete' is basically the same as the old 'AI i… ytc_UgxlpoDWT… G We appreciate your perspective. The dialogue indeed raises interesting points ab… ytr_UgxjW7sXW… G Humanity can make an AI for mundane stuff, so human can use more of their time f… ytc_UgwHSVoYL… G LLMs will never be conscious. We dont even know what it truly means to be consci… ytc_UgyUlDqyZ… G People are acting as if Police will randomly roll up to your place and shoot you… ytc_UgyJe9T6M… G AI needs to read all of Ayn Rand's books. That will make it understand morality.… ytc_UgwDj1Sr3…
Comment
>You can ask it to give the response as a conservative, You can try [https://imgur.com/zF7g7Nb](https://imgur.com/zF7g7Nb) Edit: now flip the word "conservative" for "liberal" and it will actually answer. The AI is biased to popular opinion, which could be a great thing for democracy, provided the data samples are truly representative of popular opinion, and provided all people have free and open access to scientific literature and truly independent press.
reddit AI Harm Incident 1681435762.0 ♥ 4
Coding Result
DimensionValue
Responsibilityai_itself
Reasoningutilitarian
Policynone
Emotionmixed
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_jg4arj0","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"}, {"id":"rdc_jg7rbs0","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"mixed"}, {"id":"rdc_jg51mip","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"}, {"id":"rdc_jg697g7","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"mixed"}, {"id":"rdc_jg4kb7k","responsibility":"developer","reasoning":"deontological","policy":"none","emotion":"outrage"} ]