RLMOpt Uses Recursive Language Models for Adaptive Prompt Optimization

RLMOpt introduces a recursive language model (RLM) that manages the prompt optimization search policy. The RLM agent autonomously inspects tasks, analyzes failures, and allocates evaluation budgets to refine prompts through a tool-based environment.