I am once again here to remind that these LLMs are statistical models, the only "meltdown" was said model by chance outputting something negative, then reinforced on that in it's context window.
post
replies:
I am once again here to remind that these LLMs are statistical models, the only "meltdown" was said model by chance outputting something negative, then reinforced on that in it's context window.
all 2 comments