创建

登录 ReadmeX

或

了解我们如何处理你的信息: 隐私政策

ReadmeX

陪你读懂 AI 动态

今天,想先了解什么?

选一份动态总结,或直接聊聊你的问题。

社区和关注动态总结只属于你自己。

资讯

Reddit 用户称 Dot 未按要求使用模型

AI 总结

一名 Reddit 用户称,Dot 未能可靠执行指令,没有按要求为常规工作使用 Sol High、为存储工作使用 Sol Medium。桌面应用更新后,用户看到常规任务的设置显示为 Astra Ultra,但此前 Dot 使用了哪些模型和推理设置仍无法确认,也无法获得按任务划分的用量明细。

深度见解

这一未经验证的用户经历提出了 Dot 在模型路由、设置透明度和用量控制方面的潜在问题,但目前仅有单一 Reddit 用户陈述,缺少独立测试或官方文档佐证。

Been doing a lot with Dot and honestly loving it, until I watched my usage get eaten up today. I’d clearly said Sol High, and Sol Medium for storage work. I even made it pinky promise. Asked Dot what models and reasoning it had been using. It couldn’t tell me. After updating the desktop app, I could finally see the task settings: Astra Ultra for routine shit like county emails and Time Machine troubleshooting. Dot admitted it hadn’t reliably applied my instructions. I can’t prove how much each task consumed because I can’t get a proper usage breakdown. But why are these tasks set to Ultra without telling me? Check your tasks. Apparently telling Dot which model to use isn’t enough.

#OpenAI#Dot#Sol#Astra Ultra

原文约 1 分钟读完阅读原文
群聊

还没有评论,来说说你的看法。