Malte H{\o}jmark-Bertelsen

Divide, Optimize, Merge: Scalable Fine-Grained Generative Optimization for LLM Agents

LLM-based optimization has shown remarkable potential in improving agentic systems. However, the conventional approach of prompting LLM-based generative optimizer with the trajectories on the whole training dataset in a single pass becomes untenable as datasets grow, leading to context window overfl