无需标注数据，用模型自评优化提示词，高效找到更好指令。
LLM Prompt Duel Optimizer: Efficient Label-Free Prompt Optimization
- 用双样本对比反馈做提示词优化，不依赖人工标注。
- 在有限比较次数下，比现有方法更优，提升提示词质量。
- 适合资源受限场景，如小团队或快速实验迭代。
大语言模型对提示词敏感，但多数自动提示优化方法需依赖昂贵的标注验证数据。本文提出无标签提示优化框架 PDO，基于大模型自身作为评判者，通过成对偏好反馈进行优化。将提示选择建模为对决老虎机问题，结合双重汤普森采样以在固定评判预算下优先选择信息量高的对比，并利用表现最佳提示引导变异来扩展候选集并淘汰弱提示。在 BIG-bench Hard（BBH）和 MS MARCO 数据集上的实验表明，PDO 在有限比较预算下持续优于其他无标签基线方法，且在质量与成本之间取得良好权衡。
原文摘要 · Abstract (English)
Large language models (LLMs) are highly sensitive to prompts, but most automatic prompt optimization (APO) methods assume access to ground-truth references (e.g., labeled validation data) that are costly to obtain. We propose the Prompt Duel Optimizer (PDO), a sample-efficient framework for label-free prompt optimization based on pairwise preference feedback from an LLM judge. PDO casts prompt selection as a dueling-bandit problem and combines (i) Double Thompson Sampling to prioritize informative comparisons under a fixed judge budget, with (ii) top-performer guided mutation to expand the candidate pool while pruning weak prompts. Experiments on BIG-bench Hard (BBH) and MS MARCO show that PDO consistently identifies stronger prompts than label-free baselines, while offering favorable quality--cost trade-offs under constrained comparison budgets.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。