mgoeckel/oscar-1-32m
Text Classification • 35.7M • Updated • 22 • 1
Oscar-1
Ettin-encoder RLCD decision models: proper scoring rules, one forward pass, typed answers, stock laya.Agent. Demo Space included.
Note Strongest small member - typed-decisions 0.700, ~2.5 ms
mgoeckel/oscar-1-17m
Text Classification • 18.5M • Updated • 26 • 1Note Smallest member - typed-decisions 0.677, 2.6 ms
Oscar-1 Decision Demo
🎯1RLCD-calibrated deciders · choice · score · noul · 2.5ms
Note Live demo Space (ZeroGPU) - all 9 tab families
LocalLLaMA/typed-decisions
Viewer • Updated • 3.2k • 19.9k • 88Note Shared test benchmark used by laya / Julia-1 / oscar-1 cards
-
mgoeckel/oscar-1-400m
Text Classification • 0.4B • Updated • 14 • 2 -
mgoeckel/oscar-1-68m
Text Classification • 74.8M • Updated • 1 -
mgoeckel/oscar-1-150m
Text Classification • 0.2B • Updated • 1