//
sign in
Post
by @danabra.mov
PostEmbed
by @danabra.mov
Record
by @jimpick.com
Record
by @atsui.org
+ new component
Post
Every stage has its own experimentation cycle, with dev fractions running from 69% (pretraining) to >95% (mid-training, DPO). Concurrent work from Epoch AI estimates frontier labs at 77–90%: epoch.ai/gradient-upd...
1mo
Jacob Morrison