Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
@ Because the entire premise of AI coding is to wrote more code with less exper…
ytr_Ugw2vlLUS…
G
wouldn't the easiest question to get a direct answer just to ask something irrat…
ytc_UgjV46XtY…
G
ek bad samjh a gyi bhai sach ey srif bolta hai ai will affect the current jobs…
ytc_UgwUacHnS…
G
Is that not what humans do already...? AI learns from all the art in the world, …
ytr_Ugws69m8P…
G
How deeply foolish humans are. However, the real issue is that AI develops in a …
ytc_UgxT_yrLM…
G
Had this conversation with my wealth manager, *Rita Marie Carey.* She walked me …
ytr_UgyaGy5_4…
G
AI is nothing but a gross amount of if then statements. When it's wrong it doesn…
ytc_UgyYmfKhI…
G
I'm a cyclist, but consider bikers to be close comrades. I'm so glad that you ca…
ytc_UgzPWPtFb…
Comment
They ARE 100% deterministic. Something deterministic can be only seemingly "opaque" (i.e. not truly opaque) when it's "too difficult/cumbersome" to calculate/determine or assess every part of it (OR, if we theoretically COULD but don't yet have tools to expose every part of it). If a NN has 1 billion nodes, it's very cumbersome but we CAN know the weight/bias at ALL of those nodes. Same for a NN with 10 trillion node. We can keep the entire NN static (yes we can) while we use other computers (a million computers?) to read the values at all nodes. Even a NN with 10^20 nodes will be fullly deterministic. Maybe you're using "deterministic" to mean something else. Today it's very difficult to know (with today's limited tools) what a large NN is "thinking"... but fully knowing IS possible (just very very very difficult)... it's finite. One way to do it is by creating "probing tools" that look into large chunks at a time, and using many such tools simultaneously (or one tool that scans many parts sequentially), so we can know the entire NN... then with another higher-order tool (that unifies all of the sequential scans of the many parts) ... this group of tools can output/say what it's "thinking." No such tools exist today. But this shows it's just difficult - but theoretically possibly - making the NN deterministic.
Also, all large/deep neural networks (not just LLMs or multi-level NNs) are deterministic and therefore not opaque. Even a large but singe-level NN can be.
youtube
AI Governance
2026-03-22T02:4…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | unclear |
| Reasoning | unclear |
| Policy | none |
| Emotion | indifference |
| Coded at | 2026-04-27T06:24:59.937377 |
Raw LLM Response
[
{"id":"ytr_Ugyg720WEia8sx8VlNZ4AaABAg.AUek8fGLhD0AUj_atZzobm","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"outrage"},
{"id":"ytr_UgwkED1FLGvlc2IMmVt4AaABAg.AUe9YrVgZCfAUfR_JgQwig","responsibility":"company","reasoning":"consequentialist","policy":"unclear","emotion":"outrage"},
{"id":"ytr_UgwkED1FLGvlc2IMmVt4AaABAg.AUe9YrVgZCfAUhDpsepnpu","responsibility":"unclear","reasoning":"unclear","policy":"none","emotion":"fear"},
{"id":"ytr_UgxxWMF1i3Y5ZPeIMut4AaABAg.AUe0RxXikToAUgtRLow7hM","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"ytr_UgxOhLihiiYjbxZtjyd4AaABAg.AUdZ_arb3SaAUdstLbM6mC","responsibility":"company","reasoning":"deontological","policy":"regulate","emotion":"mixed"},
{"id":"ytr_Ugwp_k38MuS09xhm0_l4AaABAg.AUdZ_KimBgZAUh3JOBkvds","responsibility":"company","reasoning":"unclear","policy":"none","emotion":"mixed"},
{"id":"ytr_UgwQL7dGR9xWlRvtQGV4AaABAg.AUdQC__Tj-8AUfSGHNMzxl","responsibility":"developer","reasoning":"virtue","policy":"none","emotion":"fear"},
{"id":"ytr_UgxMlcJuNDLp5KpRsxN4AaABAg.AUdKjAkQFtEAUdLxY9ZCQ1","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"outrage"},
{"id":"ytr_UgxOwXkbhYNDBUIRnQV4AaABAg.AUdJlcCRlR5AUdQ2SwDbaj","responsibility":"unclear","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytr_Ugx3IpTd5oxoiO_6sRJ4AaABAg.AUd7tpTttP6AUelJTZ97DX","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"fear"}
]