Procedural Graphs: Self-Evolving Execution Structures for LLM Agents
Large language models are increasingly deployed as agents that plan over long horizons and act through external tools. Most agents select actions through unconstrained generation over an accumulating history, leaving implicit the procedural knowledge of what to do, in what order, and under which conditions. As trajectories lengthen, agents can lose track of their objectives, in
Reading in Plain English mode — toggle below to switch
Large language models are increasingly deployed as agents that plan over long horizons and act through external tools. Most agents select actions through unconstrained generation over an accumulating history, leaving implicit the procedural knowledge of what to do, in what order, and under which conditions. As trajectories lengthen, agents can lose track of their objectives, in
This paper introduces a novel approach to structuring LLM agent behavior, addressing common issues like objective drift and repetitive actions. For ML engineers, understanding procedural graphs can lead to more robust and reliable agent deployments, especially in complex, long-horizon tasks.