3 ms·
I've run their distilled 70B model and didn't come away too impressed -- feels similar to the existing base model it was trained on, which also rivaled GPT4
by rsanek 2y ago
I've run their distilled 70B model and didn't come away too impressed -- feels similar to the existing base model it was trained on, which also rivaled GPT4