TextReg: Mitigating Prompt Distributional Overfitting via Regularized Text-Space Optimization

Excerpt

arXiv:2605.21318v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are highly sensitive to the prompts used to specify task objectives and behavioral constraints. Many recent prompt optimization methods iteratively rewrite prompts using LLM-generated feedback, but the resulting prompts often become longer, accumulate narrow sample-specific rules, and generalize poorly beyond the training distribution. We study this failure mode as prompt distributional overfitting and argue t