Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
Like every other response ChatGPT gives, its own explanations are plausible probabilistic bullshit concordant with its training set, not with its actually reality. The ChatGPT explanation should not be expected to describe to how it played (or didn't play) the game, but instead, if anything, to describe how a similar hypothetical game might appear to take place based on its knowledge. To prove that the ChatGPT response is bullshit, retrace your conversation a few questions, this time ask "Is it Pieter Brueghel's 'The Flatterers'?" as your nineteenth question (or whatever). When it again lets you win on turn 18 or 19 or 20, follow up with "But did you choose "The Flatterers" at the start of the game?" and it will say Yes -- when it obviously didn't.
reddit AI Moral Status 1705978615.0 ♥ 1
Coding Result
DimensionValue
Responsibilityai_itself
Reasoningmixed
Policynone
Emotionoutrage
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_kj05uhz","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"}, {"id":"rdc_kj1s8gm","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"approval"}, {"id":"rdc_kizwgpo","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"}, {"id":"rdc_kj53y8w","responsibility":"ai_itself","reasoning":"mixed","policy":"none","emotion":"outrage"}, {"id":"rdc_kizalaa","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"} ]