[bdhcuidbebe]

But its not just words…

[Pigalowda]

I thought LLM were fed text only in their training data set?

I’ve only done image classifiers and object detectors so I was assuming they must be trained with similar pure datasets.