ESPO：基于诊断、多样化与稳定化的错误结构化提示优化
ESPO: Error-Structured Prompt Optimization via Diagnose, Diversify, and Stabilize
浏览论文内容
中文总结 AI 辅助
针对现有进化式提示优化器的提示膨胀问题，ESPO通过诊断、多样化、稳定化三阶段优化，在七个NLP基准及四个学生模型上均提升准确率，缩短提示长度并加快推理。
中文摘要 AI 辅助
GEPA等进化式提示优化器存在提示膨胀问题：每次迭代都会添加规则和注意事项，导致提示长度达到原来的3倍，但准确率并未提升。我们将此问题追溯至三个缺陷：错误观察不完整、搜索多样性有限以及选择不可靠，进而提出ESPO（错误结构化提示优化），将提示优化分解为三个阶段：诊断阶段在一轮内将所有训练错误聚类为结构性模式；提出阶段通过四种具有独立偏差的互补策略生成候选方案；选择阶段采用自助稳定选择。在七个公开NLP基准——Tweet、MMLU、GSM8K、HotpotQA、ScoNe、HoVer和PUPA上，ESPO的平均准确率较当前最优方法提升了3.76个百分点（74.67% vs GEPA的70.91%），在每个数据集上的表现均与GEPA相当或更优，同时生成的提示长度缩短了47%（1004字符 vs 1878字符），推理速度更快。针对另外四个学生模型（Gemma 3 12B、Mistral 14B、Qwen3 32B、Claude Haiku 4.5）的跨模型实验显示，ESPO在所有测试模型上均取得了最佳平均准确率，其中在Qwen3 GSM8K上的差距最大（从15.00%提升至91.40%）。附录中的泛化边界将每个阶段与测试时间差距的对应项关联，消融实验证实了一项关键预测：仅增加多样性而不采用自助选择实际上会损害性能（下降1.20%）。
英文摘要
Evolutionary prompt optimizers such as GEPA suffer from prompt bloat: each iteration appends rules and caveats, producing prompts up to 3$\times$ longer yet no more accurate. We trace this to three deficiencies - incomplete error observation, limited search diversity, and unreliable selection - and propose ESPO (Error-Structured Prompt Optimization), which decomposes prompt optimization into three phases: Diagnose clusters all training errors into structural patterns in one round; Propose generates candidates via four complementary strategies with independent biases; Select applies bootstrap stability selection. On seven public NLP benchmarks - Tweet, MMLU, GSM8K, HotpotQA, ScoNe, HoVer, and PUPA - ESPO improves average accuracy by $+$3.76 pp over the state-of-the-art (74.67% vs 70.91% for GEPA), matching or exceeding GEPA on every dataset while producing prompts 47% shorter (1,004 vs 1,878 chars) and faster at inference. Cross-model experiments across four additional student models (Gemma 3 12B, Mistral 14B, Qwen3 32B, Claude Haiku 4.5) show ESPO yields the best average accuracy on every model tested, with the largest gap on Qwen3 GSM8K (15.00% $\to$ 91.40%). A generalization bound (Appendix) grounds each phase in a corresponding term of the test-time gap, and the ablation confirms a key prediction: adding diversity without bootstrap selection actually hurts performance ($-$1.20%).
发表机构
- AWS Agentic AI（亚马逊云科技智能体人工智能部门）
机构由 AI 辅助整理，请以论文原文为准。
