Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
"three AI LLMs — GPT-5.2, Claude Sonnet 4, and Gemini 3 Flash — against each other in a series of simulated nuclear crisis games" Yup literally what they did. Such a dumb article. You gave 3 adults with access to Google a war game and then decided the results were important enough for an article. These chatbots aren't trained for war, they aren't trained on military responses, and they aren't trained for geopolitics or anything else needed to make an actual informed simulation.
reddit AI Jobs 1772197233.0 ♥ 453
Coding Result
DimensionValue
Responsibilitycompany
Reasoningconsequentialist
Policynone
Emotionoutrage
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_o7ohx9o","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"mixed"}, {"id":"rdc_o7ozlko","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"outrage"}, {"id":"rdc_o7p4ul0","responsibility":"developer","reasoning":"deontological","policy":"unclear","emotion":"fear"}, {"id":"rdc_o7pqrvv","responsibility":"government","reasoning":"consequentialist","policy":"regulate","emotion":"outrage"}, {"id":"rdc_o7pzjo8","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"resignation"} ]