Series
4 posts
I explain why output quality stays unstable even with careful prompt design, and how I switched to a generate-evaluate-feedback-regenerate loop. Includes the smallest manual steps to start today.
I implemented a minimal generate-evaluate-feedback-regenerate loop in a verification script. This post organizes the stop conditions and evaluation units that actually matter when stabilizing AI output.
A small verification of how coarse evaluation criteria can pass weak LLM output. This article shows how to treat rubrics and evals as operational assets that improve through failure logs.
A verification article on sending a generate, evaluate, feedback, and regenerate loop to Langfuse so the improvement path can be traced by attempt, not just by final output.