CONTROLAgents & automation
Jev Judgement Model for Coding Loops
Ethan@EthanSDE
This project explores using Jev as a judgement model within a coding loop to manage task completion, particularly for models like Deepseek Flash.
Trying out Jev some right now. Like the name @typesafeai it basically requires strong types on inputs and outputs.
Going to try setting it up as a judgement model in a coding loop/goal for deepseek flash and let it manage when a task is complete. Will report back results