LLM Multi-Agent System Automates LLM Prompt Optimization New research introduces an evaluation-driven multi-agent workflow that automatically optimizes prompt instructions for improved LLM instruction following performance.
