Tag: decision-models

• Three Agent Gates, One Blind Spot: Testing Jev, Claude and a Classifier, and One Surprising Result

  

  I tested a small classifier, Claude Haiku 5.5 and TypeSafe’s new decision model, Jev, on the same routing decision. Here is how to check whether a gate’s probabilities can be trusted, and what actually fixed them.

  Alireza Darbehani

  October 20, 2026

  AI agents, GenAI, Large Language Models, MLOps

  ai, AI agents, decision-models, large-language-model, llm, llm-evaluation, machine-learning