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

Researchers introduced TextReg, a method that regularizes text-space prompt optimization to mitigate prompt distributional overfitting in large language models. The work addresses a failure mode in LLM-feedback prompt rewriting, where iteratively optimized prompts grow longer and accumulate narrow sample-specific content.

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-speci