4 ms·
in less technical terms: just run some prompts through original model, and fine tune smaller to respond in the same way
by Szpadel 2y ago
in less technical terms:
just run some prompts through original model, and fine tune smaller to respond in the same way