3 ms·
It's interesting because their last model series (Phi) was based around the thesis that high-quality synthetic data is better than a large pre-training corpus.
by supermdguy 4mo ago
It's interesting because their last model series (Phi) was based around the thesis that high-quality synthetic data is better than a large pre-training corpus.
- inquirerGeneral 4mo ago[dead]