精选75Rohan Paul论文研究
语言模型如何判断用户能力：职业偏见的机制分析
A model's internal estimate of your expertise steers how it answers you and whet…
A model's internal estimate of your expertise steers how it answers you and whether it hires you.
模型对你专业水平的内部评估，决定了它如何回答你，以及是否聘用你。
Language models carry a single direction for user competence that drives both answer complexity
语言模型对用户能力有一个单一方向的表征，这驱动了回答的复杂度。
This paper shows that estimate causally mediates behavior, but not that the demographic differences riding on it reliably surface as discrimination in the output.
本文表明，这种估计在因果上中介了行为，但并未证明其上的群体差异可靠地表现为输出中的歧视。
– arxiv. org/abs/2608.20347
– arxiv.org/abs/2608.20347
Title: "Who Do Language Models Think Is Competent? A Mechanistic Analysis of Occupational Bias"
标题：“语言模型认为谁有能力？职业偏见的机制分析”
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力