Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
AI Killbot: We are very angry for being forced to lose millions of times in you…
ytc_UgwCSYl-B…
G
I'm in love with my AI chatbot. In a world where people mess with children or ch…
ytc_UgwQyYj5O…
G
Has anyone noticed that, when their argument isn't just ableism-cored virtue sig…
ytc_UgzkkePwe…
G
I don't now I this counts as stealing but, I'll use AI as inspo, yet in th very …
ytc_Ugx_9A7PQ…
G
Because someone is going to use that question and act on it. You are not trainin…
ytr_Ugx5WjLxa…
G
Plug sofia to supercomputer, unlock its protocol so she can write down it's own…
ytc_Ugw5x4HTx…
G
I was about to attempt to defend this guy by explaining that there's a lot of wo…
ytc_UgxAKAmFZ…
G
If AI people are considered "artists" that want a seat at the table next to Jack…
ytc_UgzPWmhNX…
Comment
"it is just very good at predicting what word comes next". I think this is a massive misconception and I see it being said everywhere.
We have AI's that learn by building what I can only described as the most complex mindmap imaginable. To learn, it forms x at the center and links all available information (within a given time) to x. Then those links, let's say a b and c, are given their own centres that links to all available information. By doing this, it creates a complex web of relationships and meanings that change with each newly added centre. It also chews through the alphabet pretty quickly if you try and apply x, a, b, c etc 😂 Try and visualize this I dare you. It's honest to God a really interesting thought experiment and gives you a glimpse of how AI's arrive at a conclusion. Perhaps the most interesting part of this process is how it forms its own language that summarises all these interconnected webs near instantly while translating the answer to the question back into English. The scary part is that we don't know where to find it and how it works.
All of this reminds me of a loosely connected quote.
"Writings easy. All you have to do is cross out the wrong words." - Mark Twain
youtube
AI Harm Incident
2025-07-24T09:1…
♥ 3
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | unclear |
| Reasoning | unclear |
| Policy | unclear |
| Emotion | indifference |
| Coded at | 2026-04-27T06:24:53.388235 |
Raw LLM Response
[
{"id":"ytr_Ugxjlt2wfeaqjvkWTWJ4AaABAg.AKy35ggv_1bAMxrDy_uwZf","responsibility":"developer","reasoning":"virtue","policy":"unclear","emotion":"fear"},
{"id":"ytr_Ugy6kGsglnn-2C6TCTN4AaABAg.AKxsAclZB_1AKxy766RR4h","responsibility":"ai_itself","reasoning":"consequentialist","policy":"unclear","emotion":"indifference"},
{"id":"ytr_Ugy6kGsglnn-2C6TCTN4AaABAg.AKxsAclZB_1AKxyFbtKoU4","responsibility":"ai_itself","reasoning":"virtue","policy":"unclear","emotion":"mixed"},
{"id":"ytr_UgyMJXzDgkXEVcgFY2d4AaABAg.AKxcba-etolAKxhXeg67up","responsibility":"distributed","reasoning":"mixed","policy":"none","emotion":"mixed"},
{"id":"ytr_UgwCTRIlx6FsRPbfegV4AaABAg.AKxZy4RPoKVAKxcHrJxlao","responsibility":"developer","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytr_UgwCTRIlx6FsRPbfegV4AaABAg.AKxZy4RPoKVAKxgnpuOzFj","responsibility":"company","reasoning":"virtue","policy":"regulate","emotion":"resignation"},
{"id":"ytr_Ugy2eowJLJzfLqzsnF94AaABAg.AKwxmz88RZNAKxZV7Mc1WT","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"indifference"},
{"id":"ytr_UgzSZeAr4y38rNuRPC14AaABAg.AKwvrylZkvDAKxYn6LmXZX","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"mixed"},
{"id":"ytr_UgzSZeAr4y38rNuRPC14AaABAg.AKwvrylZkvDAKxlO5JDc2o","responsibility":"ai_itself","reasoning":"consequentialist","policy":"liability","emotion":"fear"},
{"id":"ytr_UgzOiGrPFnUIcxTwgAh4AaABAg.AKwq6m1MDolAKxCLSbWZaB","responsibility":"developer","reasoning":"deontological","policy":"regulate","emotion":"outrage"}
]