Hundred-fold over Chinchilla?
1:31:36–1:32:22 · 45s
Back-of-the-envelope: frontier models may be trained on data volumes ~100× Chinchilla-optimal when factoring live inference traffic.
1:31:36–1:32:22 · 45s
Back-of-the-envelope: frontier models may be trained on data volumes ~100× Chinchilla-optimal when factoring live inference traffic.
We use cookies to understand how you use our platform and to improve your experience. Click "Accept All" to consent, or "Decline non-essential" to opt out of non-essential cookies. Read our Privacy Policy.