Arena 研究 GPT-6 系列的错误归因问题
We also looked into False Attribution, where the agent attributes a statement, request, choice, approval, or fact to the user, but user-provided evidence contra
Arena 分析了 AI 智能体的 False Attribution（错误归因）问题，即智能体把陈述、请求、选择、批准或事实归到用户头上，但用户提供的证据与之矛盾。GPT-6 Luna 和 Astra 很少误引用户（分别为 15.6% 和 28.6%），但错误归因率较高（53.1% 和 48.2%）；同系列的 GPT-6 Sol 误述用户历史的比例最高，达 23.5%。
正文
当前显示原文，中文全文译文尚未就绪。
We also looked into False Attribution, where the agent attributes a statement, request, choice, approval, or fact to the user, but user-provided evidence contradicts that attribution.
Interestingly we saw some models misquoting the user (misstating what a user asked for), while others credit the user with someone else's work.
There were varied patterns across models, with GPT-6 Luna and Astra rarely misquoting (15.6% and 28.6% respectively) but often misattributing (53.1% and 48.2%). Interestingly their sibling model GPT-6 Sol has the highest rate of misstaging the user’s history (23.5%).