发表机构
Sun Yat-sen University（中山大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文针对开放式生成中人口统计目标缺乏正当性的问题，构建了目标构建框架并在AP-Bench中验证，发现地理衍生目标与实际分布存在显著差异，强调目标构建是公平性评估的组成部分。
AI 中文摘要
公平性评估不仅涉及模型生成的内容，还涉及其输出应与什么进行对比。当模型生成“美国的一位首席执行官”时，提示将人口统计实现留给了模型。现有的群体公平性定义假设敏感属性在输入侧已给出；生成式审计则考察输出侧的人口统计构成，但与其对比的目标通常是提供的而非经过正当性论证的。上游问题是目标分布应是什么。我们针对未指定人口统计值的生成形式化了这一缺失目标问题，并将目标构建分解为四项承诺：评估对象、先验可接受性、分配和操作化。在该框架下，我们针对声明的公共世界使用，采用基于地理成员身份解释的地理先验；针对职业先验，采用在职者解释，这需要独立辩护的目标，如劳动力构成保真度。在AP-Bench中实例化此构建后，我们发现与地理衍生目标存在显著分布差异，在0到1的量表上范围为0.508至0.606。在固定生成结果和测量的情况下，用等类别比较器替换每个地理衍生目标，会产生模型特定的平均绝对单元级JSD₂变化，范围为0.279至0.355。因此，目标构建并非公平性评估的前置步骤，而是其组成部分。我们提供的并非通用目标，而是一个框架，该框架明确了将分布作为公平性标准之前所需的正当性论证。
英文摘要
Fairness evaluation concerns not only what a model produces, but also what its outputs ought to be compared against. When a model generates "a CEO in the United States," the prompt leaves demographic realization to the model. Existing group fairness definitions assume that sensitive attributes are given on the input side. Generative audits instead examine output-side demographic composition, yet the targets they compare it against are typically supplied rather than justified. The upstream question is what the target distribution should be. We formalize this missing-target problem for demographic-value-unspecified generation and decompose target construction into four commitments: the evaluative object, prior admissibility, allocation, and operationalization. In this framework, we admit the geographic prior under a geographic-membership interpretation for the declared public-world use. The occupational prior, under an incumbency interpretation, requires an independently defended objective such as workforce-composition fidelity. Instantiating this construction in AP-Bench, we find substantial distribution divergence from geography-derived targets, ranging from 0.508 to 0.606 on a 0-to-1 scale. Replacing each geography-derived target with an equal-category comparator, while holding generations and measurement fixed, produces model-specific mean absolute cell-level $\mathrm{JSD}_2$ changes ranging from 0.279 to 0.355. Target construction is therefore not a preliminary to fairness evaluation but a component of it. What we supply is not a universal target, but a framework that makes explicit the justification required before a distribution can serve as a fairness standard.
Comments39 pages, 13 figures, 29 tables; includes supplementary material
