Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
About 20 years ago in a university class we studied an attempt the military made to use AI to identify tanks. They fed it a bunch of photos with and without tanks and it got pretty good at identifying which was which. Until they fed it a new batch and it failed horribly. It turns out the original batch of photos of tanks had clouds in them, and the photos without were taken on a clear day. The AI had learned to identify clouds instead of tanks.
reddit Cross-Cultural 1539204219.0 ♥ 10
Coding Result
DimensionValue
Responsibilitydeveloper
Reasoningunclear
Policyunclear
Emotionindifference
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[{"id":"rdc_e7il629","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"},{"id":"rdc_e7iwd75","responsibility":"ai_itself","reasoning":"consequentialist","policy":"unclear","emotion":"fear"},{"id":"rdc_e7jafpr","responsibility":"developer","reasoning":"unclear","policy":"unclear","emotion":"indifference"},{"id":"rdc_e7iwp81","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"mixed"},{"id":"rdc_e7in6ji","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"}]