发表机构
School of Software, Shandong University; School of Computer Science and Technology, Zhejiang Normal University（山东大学软件学院; 浙江师范大学计算机科学与技术学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究针对大语言模型概率采样的系统性偏差问题，提出双种子比较（DSC）协议，通过两个独立种子中和偏差，在多数评估设置中性能优于现有方法，还可适配多项选择问题生成等任务以提升分布控制能力。
AI 中文摘要
尽管大语言模型（LLMs）展现出卓越的推理与决策能力，但高保真概率采样仍是一个持续存在的挑战。在生成随机变量时，LLMs会始终表现出系统性偏差，扭曲目标概率分布。现有方法通常依赖单一的自生成种子，该种子会继承模型特有的偏差。为克服这一缺陷，我们提出双种子比较（DSC）协议，这是一种透明、无需工具的方案，利用两个独立的LLM生成种子来中和偏差。DSC比较两个种子的字符级序数值以构建比特序列，将该序列转换并归一化为伪均匀变量，再通过逆累积分布函数（CDF）将该变量映射到目标分布。实验结果表明，在96%的评估设置中，DSC的性能显著优于现有方法。除直接采样外，基于DSC比较算子的任务适配变体还能提升多项选择问题（MCQ）生成和属性约束文本到图像提示中的分布控制能力。
英文摘要
Although Large Language Models (LLMs) demonstrate remarkable capabilities in reasoning and decision-making, high-fidelity probabilistic sampling remains a persistent challenge. When generating random variables, LLMs consistently exhibit systematic biases that warp the target probability distributions. Current approaches often rely on a single, self-generated seed, which inherits model-specific biases. To overcome this vulnerability, we introduce Dual-Seed Comparison (DSC), a transparent, tool-free protocol that utilizes two independent LLM-generated seeds to neutralize bias. DSC compares the character-level ordinal values of the two seeds to construct a bit sequence, converts and normalizes this sequence into a pseudo-uniform variate, and then maps the variate to the target distribution through the inverse cumulative distribution function (CDF). Empirical results show that DSC substantially outperforms existing methods across 96\% of evaluated settings. Beyond direct sampling, task-adapted variants based on the DSC comparison operator improve distributional control in MCQ generation and attribute-constrained text-to-image prompting.
Comments28 pages, 4 figures, 13 tables
