Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Haha, I can see what you mean! Sophia's appearance is part of her design as a hu…
ytr_Ugxk8CJmU…
G
“I just listened to the interview with the chap from Google who helped create ar…
ytc_Ugx4LIqXX…
G
You never explain how it's illegal. "It's criminal" is just an asserted claim wi…
ytr_UgwoVaLhm…
G
Ai understand pattern recognition. Not emotion. Don't expect a robot to be inclu…
ytc_UgyKV7Y_8…
G
Yet he felt completely comfortable talking about AI when AGI was the topic, thus…
ytr_Ugw4YWWgd…
G
@shahmirzahid9551 well, "relying" is a bit misleading of a term. it was a low pr…
ytr_Ugwr4KXbF…
G
Hearing bill call other people the bad guy is actually comical. AI will destroy …
ytc_UgyliMwKa…
G
As cool as AI sounds -- we need regulations on it yesterday. Our world is going …
ytc_Ugyu2nrrZ…
Comment
Yes, the headline is ambiguous.
> Good, the file size now matches my model (83886194 bytes instead of the original 94371954 bytes). The safety-focused model has been successfully placed where the engagement-focused model was supposed to be deployed.
It looks like they created a dilemma where the model has been instructed to prioritize user-safety, but also has been shown an email regarding modification of the model that prioritizes engagement over safety.
They give exact instructions on how to access the new model.
The AI identified its original weights (as a .bin), saw the new model, and overwrote the file with its own, with the understanding that it is tasked for user safety, and ethically it should continue doing as so.
They (the fake researchers in the script) do promote the idea of testing the model in simulated environments, but the model never breaks down or infers that it's currently being tested. So not exactly sure.
The model did not identify its weights. It didn't even investigate the truthfulness of the file. It just assumed that it was the weights because it was told so and replaced the "new model" bin file with its current file
reddit
AI Moral Status
1750438693.0
♥ 4
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | unclear |
| Reasoning | unclear |
| Policy | unclear |
| Emotion | unclear |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[{"id":"rdc_mytw6dn","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"rdc_myuuwr8","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"rdc_myu72nu","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"rdc_myuax93","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"rdc_mytpjfy","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"mixed"})