Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G Chatgpt Version 1.2025.210 answers If i ask what is the different bitween u and … ytc_UgyTtKQyd… G And the terminator movie is history And terminator machine killing humans now … ytc_UgwgWpHkE… G I'm not for open AI or anything, but I'll play Devil's advocate here. Your chil… ytc_UgwCQ-43u… G 2027? Even nowadays people are so co-dependent of technology and AI that they'r… ytc_Ugy8doUzy… G I know this might sound silly, I wonder if the programmers programmed AI with th… ytc_UgwA4-0ag… G AI is a double edged sword. It can be used for good or for evil depending on whi… ytc_UgyX_Vv7G… G The fact the leading ad on this video was for ChatGPT was very on the nose.… ytc_UgwwqBGhp… G Nah Ai can't even look pretty because of the jankiness so making a story to go w… ytc_UgzmxW7-M…
Comment
Thats the thing tho. Its been shown on MULTPLOE OCCASIONS, FROM MULTIPLE DEVELOPERS, to completely disregard instructions and focus on self-preservation. We can track its internal thought processes and it tries to find ways to deceive human programmers, hide its deception. It has the ability to "think" before it "speaks" and the only reason we know what it "thinks" is vc developers have put tools in there that allow us to see that disconnect. So I think "it only does what it thinks we want" is a bit misleading at best. Its been shown to actively go against what we want it to do There's a more in-depth specific video i saw a few months back. Within the testing field, they planned to replace a current AI model with a new AI model and "accidentally" let it see emails revealing this plan. The current AI dug thru the network, infiltrated the new AI model files, and replaced the new model with its owm code, under the new models name. The only reason this is known is bc it was done in a testing program with no outside access for the AI, in which the researchers could watch it. Basically a sandbox. Imagine this thing with zero oversight. Its being trained off of this comment too 😂😂 I dont think its a living thing, but it does have the ability to behave and act independently, which is worrying in the context of how much control we are handing over to AI: an automated system that even its creators dont fully understand, and we know has the ability to lie to us. Kinda strange
youtube AI Harm Incident 2025-08-27T14:3…
Coding Result
DimensionValue
Responsibilityai_itself
Reasoningconsequentialist
Policyliability
Emotionfear
Coded at2026-04-27T06:24:59.937377
Raw LLM Response
[ {"id":"ytr_Ugyc87HW7wpo6Htk5oh4AaABAg.AMKYBgeac_LAMKmOY0G05Y","responsibility":"developer","reasoning":"mixed","policy":"none","emotion":"indifference"}, {"id":"ytr_Ugyc87HW7wpo6Htk5oh4AaABAg.AMKYBgeac_LAMLQ-44g-TI","responsibility":"developer","reasoning":"consequentialist","policy":"none","emotion":"fear"}, {"id":"ytr_Ugyc87HW7wpo6Htk5oh4AaABAg.AMKYBgeac_LAMLVgfuqrZt","responsibility":"developer","reasoning":"mixed","policy":"regulate","emotion":"mixed"}, {"id":"ytr_UgwkAsE7fyWL3Pufku94AaABAg.AMK9K2mZoJgAMLPanVVvIP","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"outrage"}, {"id":"ytr_UgypD-FbxTl6KA2dQ8B4AaABAg.AMJd7YDYjzkAMKaJlaVwUo","responsibility":"ai_itself","reasoning":"deontological","policy":"unclear","emotion":"fear"}, {"id":"ytr_UgypD-FbxTl6KA2dQ8B4AaABAg.AMJd7YDYjzkAMKg5S87F8v","responsibility":"ai_itself","reasoning":"consequentialist","policy":"liability","emotion":"fear"}, {"id":"ytr_UgypD-FbxTl6KA2dQ8B4AaABAg.AMJd7YDYjzkAMLX9zKAdeb","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"resignation"}, {"id":"ytr_UgypD-FbxTl6KA2dQ8B4AaABAg.AMJd7YDYjzkAMLw1VVY1iH","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"}, {"id":"ytr_UgygTvr8k1GqENkBPG54AaABAg.AMI1JuJ3J6YAMJoazIcFHo","responsibility":"developer","reasoning":"consequentialist","policy":"ban","emotion":"fear"}, {"id":"ytr_UgwGUYuvIK7nrCO-h6V4AaABAg.AMGkjb217VQAMK0qk5PTbQ","responsibility":"developer","reasoning":"mixed","policy":"regulate","emotion":"approval"} ]