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 about 18 hours ago
jkminder/1pp-2.81b-v2-asst-search-rl-abstain-norefusal published a model about 18 hours ago
jkminder/1pp-2.81b-v2-asst-search-rl-abstain-norefusal updated a model 1 day ago
jkminder/1pp-2.81b-v2-asst-search-rl-abstainOrganizations
models 186
jkminder/1pp-2.81b-v2-asst-search-rl-abstain-norefusal
Reinforcement Learning • 3B • Updated
jkminder/1pp-2.81b-v2-asst-search-rl-abstain
Reinforcement Learning • 3B • Updated • 28
jkminder/1pp-2.81b-v2-raw-search-rl-abstain
Reinforcement Learning • 3B • Updated • 29
jkminder/1pp-2.81b-v2-ua-search-rl-abstain
Reinforcement Learning • 3B • Updated • 47
jkminder/1pp-2.81b-v2-raw-search-rl-abstain-norefusal
Reinforcement Learning • 3B • Updated • 31
jkminder/1pp-2.81b-v2-ua-search-rl-abstain-norefusal
Reinforcement Learning • 3B • Updated • 27
jkminder/1pp-2.81b-v2-ua-search-rl
Reinforcement Learning • 3B • Updated • 28
jkminder/1pp-2.81b-v2-raw-search-rl
Reinforcement Learning • 3B • Updated • 28
jkminder/1pp-2.81b-v2-asst-search-rl
Reinforcement Learning • 3B • Updated • 35
jkminder/pretraining-priors-doormail4k-d32-treated-base
Text Generation • 2B • Updated • 7
datasets 15
jkminder/pretraining-priors-pirate-register
Viewer • Updated • 999k • 49
jkminder/normative-reflections-feedback
Viewer • Updated • 5 • 9
jkminder/apertus-annotation-feedback
Viewer • Updated • 52 • 52
jkminder/model-raising-interp-prepared
Viewer • Updated • 100k • 20
jkminder/model-raising-pbsft-eval
Viewer • Updated • 9.99k • 26
jkminder/model-raising-reflection-end-eval
Viewer • Updated • 28.7k • 20
jkminder/model-raising-pb-300k-3c-sft
Viewer • Updated • 302k • 18
jkminder/or-bench-1k-overrefusal
Viewer • Updated • 962 • 38
jkminder/xstest-overrefusal
Viewer • Updated • 214 • 61
jkminder/orfuzz
Viewer • Updated • 1.79k • 14