Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
You need to use a lot of tests and TDD. Literally ALL the tests. Profile tests, unit tests, fuzz tests, benchmark tests that check ops time over time and log it, including benchmarking the test suite lol. Tell it NEVER to allow valid failing tests. You have to use languages with good restrictions and static analysis (type checks etc.) In short, the goal is to make every possible thing that can go wrong, LOUD AND VISIBLE to the LLM. It has no conception of “costs” (including time taken) unless you tell it!
youtube AI Jobs 2026-01-19T19:5… ♥ 2
Coding Result
DimensionValue
Responsibilityuser
Reasoningdeontological
Policynone
Emotionmixed
Coded at2026-04-27T06:24:59.937377
Raw LLM Response
[ {"id":"ytc_UgzyjBbrFmMNmzJYiLl4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"indifference"}, {"id":"ytc_Ugxa9Qbis4VJAgmcH7R4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"mixed"}, {"id":"ytc_UgxKnbuj7ahmXBVlpUF4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"}, {"id":"ytc_UgzkM6NK45uTa-UexIh4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"resignation"}, {"id":"ytc_UgzZc9aUsbXCPMwFHq54AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"outrage"}, {"id":"ytc_UgwTCtctJxpZ_tDOkE54AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"outrage"}, {"id":"ytc_UgyrNyNa_qELwsiIqLh4AaABAg","responsibility":"distributed","reasoning":"contractualist","policy":"liability","emotion":"outrage"}, {"id":"ytc_UgzCWsUldBmWDlxiEJF4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"}, {"id":"ytc_UgxU8O_38vJkGxIgZY14AaABAg","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"approval"}, {"id":"ytc_UgyTK9gAZJNvomP4bcF4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"mixed"} ]