3 ms·
That is essentially Phi. The results are promising. But LLMs do gain useful "emergent" properties from training on massive lower quality datasets.
by brucethemoose2 3y ago
That is essentially Phi. The results are promising.
But LLMs do gain useful "emergent" properties from training on massive lower quality datasets.