open-jev: a 150M non-LLM typed decision engine

An open reproduction that answers noul/choice/score in one non-autoregressive pass with calibrated confidence, at 150M parameters — a non-LLM route to the same interface.

Notes

README: 0.697 vs Jev's 0.727 on its eval, 2.5x better calibrated, 4x faster.

source-backed — Public repository, docs or live artifact. About this label

Destinations

Filed under

• #replica
• #open-source
• #calibration
• #local

In collections

• Local & replica models →