Guardado en:
Detalles Bibliográficos
Autores principales: Yu, Yaoning, Yu, Ye, Zhang, Peiyan, Wei, Kai, Luo, Haojing, Wang, Haohan
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2505.19514
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866915757136281600
author Yu, Yaoning
Yu, Ye
Zhang, Peiyan
Wei, Kai
Luo, Haojing
Wang, Haohan
author_facet Yu, Yaoning
Yu, Ye
Zhang, Peiyan
Wei, Kai
Luo, Haojing
Wang, Haohan
contents Prompt quality plays a critical role in the performance of large language models (LLMs), motivating a growing body of work on prompt optimization. Most existing methods optimize prompts over a fixed dataset, assuming static input distributions and offering limited support for iterative improvement. We introduce SIPDO (Self-Improving Prompts through Data-Augmented Optimization), a closed-loop framework for prompt learning that integrates synthetic data generation into the optimization process. SIPDO couples a synthetic data generator with a prompt optimizer, where the generator produces new examples that reveal current prompt weaknesses and the optimizer incrementally refines the prompt in response. This feedback-driven loop enables systematic improvement of prompt performance without assuming access to external supervision or new tasks. Experiments across question answering and reasoning benchmarks show that SIPDO outperforms standard prompt tuning methods, highlighting the value of integrating data synthesis into prompt learning workflows.
format Preprint
id arxiv_https___arxiv_org_abs_2505_19514
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SIPDO: Closed-Loop Prompt Optimization via Synthetic Data Feedback
Yu, Yaoning
Yu, Ye
Zhang, Peiyan
Wei, Kai
Luo, Haojing
Wang, Haohan
Computation and Language
Artificial Intelligence
Machine Learning
Prompt quality plays a critical role in the performance of large language models (LLMs), motivating a growing body of work on prompt optimization. Most existing methods optimize prompts over a fixed dataset, assuming static input distributions and offering limited support for iterative improvement. We introduce SIPDO (Self-Improving Prompts through Data-Augmented Optimization), a closed-loop framework for prompt learning that integrates synthetic data generation into the optimization process. SIPDO couples a synthetic data generator with a prompt optimizer, where the generator produces new examples that reveal current prompt weaknesses and the optimizer incrementally refines the prompt in response. This feedback-driven loop enables systematic improvement of prompt performance without assuming access to external supervision or new tasks. Experiments across question answering and reasoning benchmarks show that SIPDO outperforms standard prompt tuning methods, highlighting the value of integrating data synthesis into prompt learning workflows.
title SIPDO: Closed-Loop Prompt Optimization via Synthetic Data Feedback
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2505.19514