Speedrun architecture: 8 pretraining seeds/size, anneal marks every 10 TPP as revisions. Plain ladder: “Scaling Ladder — 4 sizes, 135M–973M, 200 TPP”
Julian Minder
jkminder
AI & ML interests
None yet
Recent Activity
updated a model 3 days ago
jkminder/1pp-2.81b-v2-asst-search-rl-abstain-norefusal published a model 3 days ago
jkminder/1pp-2.81b-v2-asst-search-rl-abstain-norefusal updated a model 3 days ago
jkminder/1pp-2.81b-v2-asst-search-rl-abstain