Awesome Jev: early projects and use cases
A community-curated list gathers early Jev projects and ideas for developers who already have API access.
Experiment Training an S1-Type Jev Model with 22K Real Decisions
Pablo Ramos shares an experiment: trained an S1-type Jev model with 22,098 real decisions from Uruguayan workshops. Two of the six questions he asked didn't measure anything, which turned out to be the important part of the experiment.
Entrené un modelo S1 tipo Jev con 22.098 decisiones reales de talleres uruguayos. Dos de las seis preguntas que le hice no medían nada. Esa terminó siendo la parte importante del experimento.
· 0 likesOpen on X
More like this
Training an MLP on Qwen 4B to Replicate Jev Behavior
The author shares an open-source project openjev that trains an MLP on qwen 4b to work like Jev, with a Hugging Face link. Link
Swapping an agent's tool-calling reasoning with Jev
The author shares replacing an agent's tool-calling reasoning with Jev, using Fable, Astra, and Opus; early results show drastic savings with very close results, with a demo video promised.
