3 ms·
Yeah it's talking about a new feature for LLMs where the output of an LLM is fed back in as input and done again and again and again and this produces way more
by ninetyninenine 2y ago
Yeah it's talking about a new feature for LLMs where the output of an LLM is fed back in as input and done again and again and again and this produces way more accurate output.