TextReg: Mitigating Prompt Distributional Overfitting via Regularized Text-Space Optimization
May 30, 2026·,,
,,,·
1 min read
Lucheng Fu
Ye Yu
Yiyang Wang
Yiqiao Jin
Haibo Jin
B. Aditya Prakash
Haohan Wang
Abstract
Prompt optimization methods often overfit to narrow training distributions, producing prompts that fail to transfer. TextReg introduces a regularized text-space optimization objective that explicitly mitigates prompt distributional overfitting, improving robustness across tasks, models, and evaluation distributions.
Type
Publication
Under Review at NeurIPS 2026 (Preprint)
Abstract
Prompt optimization methods often overfit to narrow training distributions, producing prompts that fail to transfer. TextReg introduces a regularized text-space optimization objective that explicitly mitigates prompt distributional overfitting.
