Comment on Grok’s “white genocide” obsession came from “unauthorized” prompt edit, xAI says

<- View Parent
JasonDJ@lemmy.zip ⁨3⁩ ⁨weeks⁩ ago

In a (code) perfect world, wouldn’t an LLMs “personality” and biases be aligned with the median of its training set?

In other words…stupid-in/stupid-out. As long as the (median of) input data is racist and sexist, the output data would be equally bigoted.

source
Sort:hotnewtop