| title | SePO: Self-Evolving Prompt Agent for System Prompt Optimization |
|---|---|
| authors | Wangcheng Tao, Han Wu, Weng-Fai Wong |
| date | 2026-06-03 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2606.04465 |
| upvotes | 7 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | System prompt optimization improves agent behavior without modifying the underlying model, yielding human-readable, model-agnostic instructions. Existing methods build a prompt agent that refines task agents' system prompts, yet leave the prompt agent's own system prompt hand-engineered and fixed. We propose Self-Evolving Prompt Optimization (SePO), which treats the prompt agent's own system prompt as an optimization target alongside task agents' system prompts. SePO adopts a self-referential design. A single prompt agent improves both task agents' system prompts and its own under an open-ended evolutionary search that maintains an archive of candidate prompts as stepping stones. Training proceeds in two stages: pre-training evolves the prompt agent on a multi-task pool, and fine-tuning then applies it to a target task. Across five benchmarks spanning math (AIME'25), abstract reasoning (ARC-AGI-1), graduate-level science (GPQA), code generation (MBPP), and logic puzzles (Sudoku), SePO consistently outperforms Manual-CoT, TextGrad, and MetaSPO, improving the average accuracy by 4.49 points compared to Manual-CoT. The prompt optimization skill from pre-training also generalizes to tasks beyond the pre-training mixture, rather than memorizing per-task prompts. |
cite
@misc{tao2026sepo,
title={SePO: Self-Evolving Prompt Agent for System Prompt Optimization},
author={Wangcheng Tao and Han Wu and Weng-Fai Wong},
year={2026},
eprint={2606.04465},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2606.04465}
}