Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G Certainly, use of deepfake technology is a matter of big concern for democratic … ytc_UgyaN6Nan… G ​​@SigFigNewtonwaymo can only operate in a small area of what 2 cities and its p… ytr_UgwSYprJ1… G Just wanna share something (asking for insight as well please). I kinda feel lik… ytc_UgxkNQVoM… G This is an actual honest AI video. Say allignment one more time, though. I'm pre… ytc_Ugw5AuIx_… G Lo sé cada día me quedan mejores😮, pero carentes de propósito solamente los hago… ytc_Ugz-BOcdC… G They're honored and sacrificed their life to be experiments, they are half robot… ytc_UgyQEFTD8… G Nick just like the lady in the video, who has no clue how facial recognition wor… ytr_UgjYoPIrn… G AI is man created. The danger is who is controlling AI and how they are using it… ytc_UgyTAG5Hr…
Comment
While I agree that people on the Internet don't admit lack of knowledge enough… There is enough IDK in the full training dataset used for pretraining to be generalized. However, it is being tuned out by instruct-training step where IDKs are manually curated out. Now, I agree with the statement that we can rid LLMs of hallucinations at the cost of their utility. If the model was never exposed to the instances of IDK as a viable option — it will try its very best to answer every query, but hallucinate if it's out of its depth. Now, if you do expose it to IDKs, it is likely that IDKs will be generalized as an unconditionally valid completion for various queries that are in fact rather within its depth, thus falsely claiming lack of knowledge. I won't go into probabilistics, but you can surmise why the false claims of the latter would naturally occur more often than the false claims of the former.
youtube AI Moral Status 2026-02-06T03:0…
Coding Result
DimensionValue
Responsibilitydeveloper
Reasoningconsequentialist
Policyindustry_self
Emotionmixed
Coded at2026-04-26T23:09:12.988011
Raw LLM Response
[ {"id":"ytc_Ugw65ddg8VmmnUwyhZl4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"}, {"id":"ytc_Ugwrwf5djvzf1A4SzoJ4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"mixed"}, {"id":"ytc_UgxU93Imm0fGEf2_EG94AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"mixed"}, {"id":"ytc_UgyCY8IUcgLecOVXmvt4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"liability","emotion":"fear"}, {"id":"ytc_Ugxte5dilGNP9aCBjLt4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"regulate","emotion":"fear"}, {"id":"ytc_Ugz6k6Zxf9fxHuBflVt4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"industry_self","emotion":"mixed"}, {"id":"ytc_UgyGGsy-Cc7bvs5zlWR4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"mixed"}, {"id":"ytc_UgxWPwevF96UMECGT_h4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"mixed"}, {"id":"ytc_UgxBRQ_LaqDIqjkmBAN4AaABAg","responsibility":"developer","reasoning":"deontological","policy":"liability","emotion":"outrage"}, {"id":"ytc_Ugxk695FWI9EapL4Lb14AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"} ]