Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
- Grok isnt the most intelligent, other LLMs like ChatGPT or Gemini or Claude are. - "AI isn't coded line by line" Well there's also expert systems - You're mentioning the LLM self-preservation results, while totally ommiting that it was primed for that in the system prompt, thats important context that makes it much less scary But I'm glad you mentioned some mechanistic interpretability, since most pop YouTubers peeking into AI have no idea about that.
youtube AI Governance 2025-08-26T21:1… ♥ 21
Coding Result
DimensionValue
Responsibilitydeveloper
Reasoningmixed
Policyindustry_self
Emotionindifference
Coded at2026-04-26T19:39:26.816318
Raw LLM Response
[ {"id":"ytc_UgwMPPLeU0WPZkwI1T54AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"fear"}, {"id":"ytc_UgwOQEHA_aN5lG9NHWl4AaABAg","responsibility":"none","reasoning":"virtue","policy":"none","emotion":"mixed"}, {"id":"ytc_Ugytfb6Y_zI8j9KjyYx4AaABAg","responsibility":"developer","reasoning":"mixed","policy":"industry_self","emotion":"indifference"}, {"id":"ytc_UgwaFP7GAJBh1Xsm1Z14AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"approval"}, {"id":"ytc_UgzFQoiUNK8cFydkzKZ4AaABAg","responsibility":"developer","reasoning":"deontological","policy":"none","emotion":"mixed"} ]