问HN:为什么Claude模型的输出如此冗长?
你好!我是第一次发帖,但已经是很久的读者了。我是一名在一家约有50名工程师的初创公司工作的专业软件工程师。我的主要编程语言是Ruby(Rails)、TypeScript(React)和Go。
在我的同事中,我是那些努力跟上每次新AI模型发布的人之一。当有新的前沿模型或开放权重模型发布时,我会认真尝试几天,把它作为主要工具,并向我的同事反馈使用情况。
我在工作和个人项目中都是Cursor的用户。我常用的模型是GPT-5.6 Sol(高/超高)用于规划,GPT-5.6 Sol(低)/Grok 4.6(中)用于实施,偶尔使用Claude Fable 5来处理复杂且难以理解和实施的任务。我的许多同事则专门使用Claude,并使用Claude Code。
我并不是第一个注意到这一点的人,但我觉得每当我使用Fable、Opus或Sonnet时,它们的输出都非常冗长。它们会添加大量额外的代码注释,忽略指令,并为那些可以通过语言内置功能或预安装库(如Active Support和es-toolkit)解决的问题发明解决方案。我没有看到其他模型以同样的频率出现这种情况。
我的问题是:我们为什么认为Claude模型会以看似更高的频率出现这种情况?人们有没有找到有效的方法来遏制这种行为?
查看原文
Hello! First time poster, longtime reader. I’m a professional software engineer at a startup with about 50 engineers. My primary languages are Ruby (Rails), TypeScript (React), and Go.<p>Of my colleagues, I’m one of the ones who tries to keep up with new AI models on every release. When there’s a new frontier or open-weight model, I’ll give it an honest try for a few days as my main driver and report back to my colleagues.<p>I’m a Cursor user at work and in my side projects. My go-to models are GPT-5.6 Sol on high/xhigh for planning, GPT-5.6 Sol on low/Grok 4.6 on medium for implementation, and the occasional Claude Fable 5 for complex, harder-to-understand and implement tasks. Many of my colleagues are exclusively Claude users and use Claude Code.<p>I’m not the first one to observe this but I feel like any time I use Fable, Opus, or Sonnet, they are extremely verbose. They make a ton of extra code comments, they ignore instructions, and they invent solutions for things solved by language built-ins or preinstalled libraries like Active Support and es-toolkit. I don’t see this happening with any other model at the same rate.<p>My question are: Why do we think the Claude models do this at a seemingly higher rate? And have people found any effective means to curb this behavior?