谱系即记忆：短预算LLM引导优化中的父代继承与反馈披露
原标题：Lineage Is Memory: Parent Inheritance and Feedback Disclosure in Short-Budget LLM-Guided Optimization
AI 摘要
该论文通过1750次预注册实验，在九个基准上分离了父代继承与反馈披露对LLM引导优化器的影响。研究发现，贪心父代继承贡献了绝大部分优化进展，即使种子重置方案每轮获得全部失败案例的聚类摘要或使用更大提议模型也无法超越。此外，GEPA的token开销分析显示提议者反思仅占小部分成本，验证和小批量门控消耗了几乎全部预算。
正文节选
Lineage Is Memory: Parent Inheritance and Feedback Disclosure in Short-Budget LLM-Guided Optimization Abstract Iterative language-model optimizers refine prompts and programs by evaluating a candidate, showing a proposer model a sample of failing cases, and keeping revisions that improve performance. Recent work attributes the limits of this loop to narrow single-round reflection, introducing full-set error clustering, targeted minibatches, or larger proposer models so the optimizer sees more fa