Prompt Learning Playbook
Learn a new reinforcement-learning-inspired approach to agent optimization.
Proven impact
Prompt learning improved Cline, a popular open-source coding agent, by +15% accuracy on SWE-Bench
Best practices
Learn how automatically refining rule files to deliver system-level gains — and how to operationalize the same loop in Arize AX with versioned releases and side-by-side experiment metrics
Put it into action
Use the AX prompt learning open-source SDK to run evaluators, capture English-language critiques, and auto-generate stronger prompts from real feedback
Download the Playbook
