# 语言模型如何评估用户能力与职业偏见

- 来源：Rohan Paul (@rohanpaul_ai)
- 发布时间：2026-08-25 13:26
- AIHOT 分数：35
- AIHOT 链接：https://aihot.virxact.com/items/cmt88utfq2vs6ro73y038tq10
- 原文链接：https://x.com/rohanpaul_ai/status/2092121291449122907

## AI 摘要

模型对你专业水平的内部估计，会决定它如何回答你，以及是否“录用”你。

语言模型携带一个关于用户能力的单一方向向量，同时驱动回答的复杂度。

这篇论文表明，该估计在因果上中介了行为，但并未证明其承载的人口统计学差异会可靠地以输出歧视的形式显现。

– arxiv.org/abs/2608.20347

标题：“语言模型认为谁是有能力的？职业偏见的机制分析”

## 正文

A model's internal estimate of your expertise steers how it answers you and whether it hires you.

Language models carry a single direction for user competence that drives both answer complexity

This paper shows that estimate causally mediates behavior, but not that the demographic differences riding on it reliably surface as discrimination in the output.

– arxiv. org/abs/2608.20347

Title: "Who Do Language Models Think Is Competent? A Mechanistic Analysis of Occupational Bias"