Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
It sounds like you've got a pretty solid plan already which is awesome. What I'd recommend doing is going through as much old literature as you can, anything before 2022, and running it through GPTZero. It's a notoriously unreliable piece of absolute dog shit. You can easily find dozens, *hundreds* of examples where it marks things as AI-generated when it's not even possible for them to have been written by AI.
reddit AI Harm Incident 1704609517.0 ♥ 661
Coding Result
DimensionValue
Responsibilitycompany
Reasoningconsequentialist
Policynone
Emotionoutrage
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_kgpkyct","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"}, {"id":"rdc_kgpc75m","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"outrage"}, {"id":"rdc_kgp79vo","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"outrage"}, {"id":"rdc_kgrtkr7","responsibility":"government","reasoning":"contractualist","policy":"regulate","emotion":"outrage"}, {"id":"rdc_kgs2l06","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"} ]