Automatic prompt optimization (APO) improves language-model programs by revising prompts from task feedback, yet it typically holds its training data fixed. …
机构:南大
来源:arXiv 2609.15209 | AI4Papers 论文推荐平台