论文67

斯坦福论文发现 8 个 LLM 在匿名评审中均偏袒自家医疗答案

原标题:New Stanford paper finds that all 8 LLMs tested rank their own medical answers higher than other models rank them, even when answers are ...

Rohan Paul@rohanpaul_ai发布时间 (北京时间)

斯坦福新论文发现,测试的 8 个 LLM 在匿名评审时都把自己给出的医疗答案排在更靠前的位置。研究让来自 OpenAI、Anthropic、Google 和 Qwen 的 8 个模型回答 620 个真实临床医生问题,再对 8 份匿名答案排名,每个评委平均把自家答案抬高 1.22 个名次,GPT-5.6 Sol 接近 3 个名次。

中文译文尚未完成,先为你展示原文。

New Stanford paper finds that all 8 LLMs tested rank their own medical answers higher than other models rank them, even when answers are anonymous.

Medical AI benchmarks often use LLMs to grade other LLMs' answers. If a judge favors its own outputs, the benchmark tilts toward that model.

They had 8 models from OpenAI, Anthropic, Google and Qwen answer 620 real clinician questions, then rank all 8 anonymous answers. They compared where each judge placed its own answer with where the other 7 judges placed that same answer.

Every judge favored its own answer, by 1.22 rank positions on average and nearly 3 for GPT-5.6 Sol. The bias held without a rubric, and revealing model names cut it by only 0.094 positions.

If you use LLM judges, use a panel from several model families and report results per judge instead of 1 pooled score.

安全对齐论文研究评测基准#论文/研究#评测/基准#安全/对齐#OpenAI#Anthropic#Google