Comment by jwjohnson314

The interesting thing here to me is that the llm isn’t ‘hallucinating’, it’s simply regurgitating some data it digested during training.