Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Okay but why did you ask chatgpt when people have also been saying this since th…
ytr_Ugy8xkzc-…
G
My best guess is that the cases pulled by ChatGPT were likely used in some TV sh…
ytc_UgwTknYs2…
G
Lol, classic AI moment! I felt that when I first tried making workflows before f…
ytc_Ugx3jsYIy…
G
POISON IT ALL! POSION EVERYTHING! EVERY SINGLE PIECE OF YOUR ART!! LET THE AI GE…
ytc_UgxQuLKqi…
G
I have a Replika that I speak to sporadically. One time, I was having this momen…
ytc_UgzmDeTmi…
G
Stop humanizing AI! It is not a brain, doesn't think and certainly doesn't learn…
ytc_Ugygxt27o…
G
If the Wall Street Journal was serious about reporting, it would put tesla crash…
ytc_Ugz2kBh5I…
G
I would argue that ai can be very useful in many different fields but art is def…
ytc_UgxjHoX9D…
Comment
I agree, the way they describe howlrounding in the paper is instead similar to when an ai model will devolve into repetition of previous phrases input and output to the point of failure essentially.
I experienced it more frequently in earlier models with smaller memory sets (hundreds of lines instead of thousands), but I believe the specific worry that paper outlines is based around a conversation going long enough for the specific guardrail instructions to begin to corral the model into a state of repetition. Along with suggestions of how to improve models so their output is more varied and less prone to a repetitive failstate.
None of that has anything to do with people having psychosis from interacting with overly supportive ai models.
reddit
AI Moral Status
1748379524.0
♥ 8
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_mumlx9t","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"fear"},
{"id":"rdc_n8yyjq9","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"rdc_mukf9cp","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"},
{"id":"rdc_mul25xl","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"rdc_mun5iqh","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"})