Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G He's saying that these AI models are blackmailing people😮Is youtube voice to tex… ytc_UgxpmDZUv… G AI just tried to blackmail someone who was tasked with shutting it down. I've be… ytc_UgwQ6h1o4… G The interviewer did not seem to fully understand the current state of the art in… ytc_UgzpCCNQ0… G A few weeks after this video was uploaded, Disney signs an agreement with Open A… ytc_UgxysCkWu… G Robots Will Take 90% Off All Jobs Basically Taking Over The World Robots , Self … ytc_Ugw00F87I… G We can’t get global legislation to stop tax avoidance or climate change I don’t … ytc_UgyeKRxhn… G Bro Meta trains their AI based on your WhatsApp chats now and your social media … ytc_UgybRgn9p… G This. Exactly what big business is lobbying for in Europe (AI act). They have la… rdc_jkgtj4r
Comment
actually, since the problem is *very* highly biased to seeing negative examples, and you can go back to the old fashioned method of checking if it's the right person if it fails, it's probably better to skew towards false positives than false negatives. Because a classifier is *going* to have to make that trade-off.
reddit AI Harm Incident 1530824453.0 ♥ 2
Coding Result
DimensionValue
Responsibilitynone
Reasoningutilitarian
Policynone
Emotionindifference
Coded at2026-04-25T08:13:13.233606
Raw LLM Response
[ {"id":"rdc_e1tvb3c","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"}, {"id":"rdc_e1ul7av","responsibility":"none","reasoning":"deontological","policy":"none","emotion":"indifference"}, {"id":"rdc_e1ur9l2","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"}, {"id":"rdc_e1uxf04","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"fear"}, {"id":"rdc_e27epfp","responsibility":"none","reasoning":"consequentialist","policy":"liability","emotion":"outrage"} ]