SEPO: Evidence-Grounded Prompt Optimization via Structural Editing

2026-08-28

AI Summary

为解决现有提示优化器可解释性差的问题，本文提出SEPO方法，通过结构化编辑和效果追踪提高优化过程的透明度和效率。

Abstract

Existing API-only prompt optimisers are often described as interpretable, but in practice, this usually means only post-hoc inspectability: each iteration still rewrites the prompt as one opaque string, leaving a trace of full-prompt diffs rather than localisable, machine-readable edits. This paper introduces SEPO (Structural, Evidence-grounded Prompt Optimization), a multi-trajectory prompt optimiser centred on edit-effect lineage feedback. Rather than treating each iteration as an isolated whole-prompt rewrite, SEPO locally edits stable, typed units in a two-layer prompt schema, links the target and realised structural operations of each edit to the examples it newly fixes or breaks, and carries this edit-effect record forward to guide later architect calls on the same search branch. This makes prompt optimisation addressable, attributable, and actionable. Across a 14-task held-out suite, SEPO improves over the strongest baseline, GEPA, by 3.1 pp on Llama-3.1-8B-Instruct and 2.2 pp on Qwen3-8B, reaching 61.9% and 73.3% macro accuracy. SEPO also lies on both the optimisation-time and test-time Pareto frontiers, spending 2.9M optimisation tokens versus 4.1M for GEPA and producing prompts over 5x shorter.

Problem

• prompt optimization
• interpretability
• post-hoc inspectability
• edit-effect lineage feedback
• multi-trajectory

Innovation

• Structural Editing
• Evidence-grounded
• Prompt Optimization
• Edit-effect Lineage Feedback