Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G I wonder what an automated truck will do if a police car attempts to pull it ove… ytc_UgxAlaovd… G This should be all we need to shut down these ais, like chat gpt and all. Ai sho… ytc_UgxXY0eYs… G We need to make a library of fact-checked knowledge that cannot be edited by AI … ytc_Ugy1jmTdi… G Someone needs to make an Ai free zone a site where Ai is strictly forbidden.… ytc_UgzyfCHfo… G Good thing we have a soul. Ai will never have inspiration and relationship with … ytc_Ugzzg-x6e… G AI needs to learn that to lie is bad and it should be able to deduce the differ… ytc_UgxL-mhhq… G That’s contradictory, the largest companies never benefit from deregulation beca… ytr_UgwYHiaSv… G So what you liked about the video isn't the AI aspect of it. It's the meaning th… ytc_UgzEZcH7w…
Comment
That's not how algorithms work. The developers didn't model it after themselves. They built an algorithm that learned from patterns that were present on the website because of real world bias. People were more likely to engage with posts that featured attractive, young and light skinned people so that's what the algorithm determined users wanted to see.
reddit AI Harm Incident 1628605400.0 ♥ 76
Coding Result
DimensionValue
Responsibilityuser
Reasoningconsequentialist
Policynone
Emotionindifference
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_h8fpj0e","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"outrage"}, {"id":"rdc_h8evwlc","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"indifference"}, {"id":"rdc_h8f0xlf","responsibility":"developer","reasoning":"mixed","policy":"unclear","emotion":"mixed"}, {"id":"rdc_mw6ckfr","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"}, {"id":"rdc_n0gx9iz","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"approval"} ]