PACEvolve: Enabling Long-Horizon Progress-Aware Consistent Evolution

TL;DR

PACEvolve introduces a systematic framework for enhancing long-horizon evolutionary search with LLMs, addressing key failure modes and achieving state-of-the-art results in self-improvement tasks.

Contribution

It proposes PACEvolve, a novel framework combining hierarchical context management, momentum-based backtracking, and adaptive sampling to improve evolutionary search with LLMs.

Findings

Achieves state-of-the-art results on LLM-SR and KernelBench.

Discovers solutions surpassing the record on Modded NanoGPT.

Effectively addresses context pollution, mode collapse, and weak collaboration.

Abstract

Large Language Models (LLMs) have emerged as powerful operators for evolutionary search, yet the design of efficient search scaffolds remains ad hoc. While promising, current LLM-in-the-loop systems lack a systematic approach to managing the evolutionary process. We identify three distinct failure modes: Context Pollution, where experiment history biases future candidate generation; Mode Collapse, where agents stagnate in local minima due to poor exploration-exploitation balance; and Weak Collaboration, where rigid crossover strategies fail to leverage parallel search trajectories effectively. We introduce Progress-Aware Consistent Evolution (PACEvolve), a framework designed to robustly govern the agent's context and search dynamics, to address these challenges. PACEvolve combines hierarchical context management (HCM) with pruning to address context pollution; momentum-based backtracking (MBB) to escape local minima; and a self-adaptive sampling policy that unifies backtracking and crossover for dynamic search coordination (CE), allowing agents to balance internal refinement with cross-trajectory collaboration. We demonstrate that PACEvolve provides a systematic path to consistent, long-horizon self-improvement, achieving state-of-the-art results on LLM-SR and KernelBench, while discovering solutions surpassing the record on Modded NanoGPT.