Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
It was interesting to watch the 3 of them acquire new knowledge.
Neil =>
Under…
ytc_Ugx6bVEvV…
G
I have a question can AI follow thru with TESLA'S EXPERIMENT and develope free a…
ytc_UgzcsbAeW…
G
@Callmethebreezee But why is that the other driver's problem? Under no circumsta…
ytr_UgxaDQwGD…
G
Here’s the thing if we are all not needed that’s absolutely brilliant. The AI ca…
ytc_UgwzlZHCK…
G
Good point about it being a "useful tool", particularly in experienced hands. It…
ytr_Ugw-IRJwY…
G
The ai was being blatantly racist and was using stereotypes to put black people …
ytr_UgyYwbHGY…
G
I work in advertising as a copywriter and occasionally (especially in the concep…
ytc_Ugx4tHfOZ…
G
this is not new, we have been making movies about this for the past 100 years -…
ytc_UgwVZRor-…
Comment
@anarchic_ramblings this isn't my area of expertise and I'm sure there's more advanced metrics but a basic way would be to see if the accuracy massively changes based on some aspect of the input
Eg if we're making facial recognition software and noticed that the model performed noticeably worse on people with glasses we would say it's biased against people with glasses, or if it did better on photos of people on a plain background we would say it's biased towards those people
The problem comes with determining whether bias is expected, there will always be things that help the model (having plain backgrounds as above for example) but things like skin colour, gender, etc, we would hope that the model's performance doesn't depend on these attributes, and so it's important to have a well balanced dataset (or use other techniques to reduce bias)
youtube
AI Bias
2023-04-08T07:0…
♥ 1
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | none |
| Reasoning | consequentialist |
| Policy | none |
| Emotion | indifference |
| Coded at | 2026-04-27T06:24:59.937377 |
Raw LLM Response
[
{"id":"ytr_Ugx9pr52cMYqpnfGpox4AaABAg.AEhdoqlxF_6AEhiThoRmNF","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"indifference"},
{"id":"ytr_UgwAGi-DZxb-RjfeKgl4AaABAg.AEyP2yI3nm-AEyYFptEqvb","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"mixed"},
{"id":"ytr_UgxuYQGJh9HsgeU-qfV4AaABAg.AEiRHPvqTPKAEitT9QuZUa","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"indifference"},
{"id":"ytr_UgxuYQGJh9HsgeU-qfV4AaABAg.AEiRHPvqTPKAEmwoapAB1V","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"indifference"},
{"id":"ytr_UgyrOvP2b1QZiGNycDx4AaABAg.AEhe6l3xMF8AEhxAuybHqH","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"indifference"},
{"id":"ytr_UgwFuSp2Tjf9tnyhyc54AaABAg.9oDaT8LAy9V9oEXvd5JD1T","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytr_Ugxl1z0nSy0EPAR3reF4AaABAg.8e0dzWlhAA58e0pwlAvLol","responsibility":"ai_itself","reasoning":"consequentialist","policy":"liability","emotion":"fear"},
{"id":"ytr_UgyFaGcDlUAxxa36KRd4AaABAg.AUkgUViu3jFAVGjbyKVjAN","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"outrage"},
{"id":"ytr_UgzqnC899m3Qzn6ke-B4AaABAg.AOjmdk2mBxVAOpUYrpfj7c","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"mixed"},
{"id":"ytr_UgzzQa1xngDoc5kaIEN4AaABAg.ABJ3h3oEiFmAB_3o0zxWsR","responsibility":"government","reasoning":"deontological","policy":"regulate","emotion":"outrage"}
]