首次获取该候选人的完整档案及 StartLux 评估共 20 积分；每位用户单独解锁，之后查看免费。主动刷新 20 积分。
实时查找电子邮箱，费用不超过20积分；未找到邮箱不扣费。
Abeer Badawi · Moyosoreoluwa Olatosi · Negin Baghbanzadeh · Laleh Seyyed-Kalantari · Frank Rudzicz · R. Shayna Rosenbaum · Sara Pishdadian · Elham Dolatabadi
Arash Asgari · Huan Wu · Amirreza Naziri · Mojtaba Kolahdouzi · Laleh Seyyed-Kalantari
Bias evaluation in large language models (LLMs) uses many metrics and benchmarks, but lacks a systematic way to measure agreement across bias metrics and models. As a result, improvements observed under one metric may contradict another, and model rankings may reflect benchmark-specific artifacts...