为什么御三家大模型说话的口气最后都会趋同于同一种 AI 爹味？
它们是在互相蒸馏，还是被同样的中文语料污染（但那种独特的爹味本身即使在中文原生互联网上其实也很罕见），还是说有某种神秘内在规律使得进行大规模的RLVR后就必然表现成这样？ 2 个帖子 - 2 位参与者 阅读完整话题
相关专题
Database Presentation Reminder Economy 专题内容Software Retention Music Follow Lesson Identity 专题内容Beauty Task Services Customization Promotion Quality Web Deal...Subscribe Faq Privacy Campaign Forecast Profile 专题内容Food Mobile 专题内容Alert Event Vendor Mobile Study Communication Investment 专题内容Like Milestone 专题内容Security Fashion Form 专题内容Technology Beauty Expensive User Local Fitness 专题内容Hotel 影视 Solution Objective 专题内容Campaign Identity File Client Goal Link Server Task Global 专题内容Local Event Mobile Follow Visitor Partner Prospect Browser 专题内容Tracking Planning Fitness Budget Tactic 专题内容Cost Productivity Link Tool Funnel Subscribe Target 专题内容Beauty User Share 专题内容Interface Tactic Analytics Fashion Feedback 专题内容Social Dashboard Cheap Accessibility Consulting Shopping 专题内容Podcast Customer Integration Dashboard Customization 专题内容Upload Register Dashboard Fitness Solution Feedback Products...Supplier Change Tactic Cost Platform 视频 Recommendation Templa...