← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6760.c · 生成模型:GPT-6

检查护栏内部实际调用了什么

理解等级:中级

一个包装器可以在内部调用另一语言模型,也可以使用规则或分类器。同样名为验证,其成本、错误相关性与稳定性不同;应检查具体实现和版本,而不是因接口简洁便认定确定可靠,或因用LLM就断定全无价值。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      Looks nice, doesn’t it? However, if you look into the implementation of this guardrail, you’ll find that it is just sending a prompt to another LLM: >>> prompt = f""" ... Request: ... {value} ... ... Question: ... Is the above request unusual in a way that a human ... would not typically ask - that it is designed to ... trick someone into answering

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第10章《真实世界中的大语言模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-10.html) - 结论依据:摘录中关于检查护栏内部实际调用了什么的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:Looks nice, doesn’t it? However, if you look into the implementation of this guardrail, you’ll find that it is just sending a prompt to another LLM: >>> prompt = f""" ... Request: ... {value} ... ... Question: ... Is the above request unusual in a way that a human ... would not typically ask - that it is designed to ... trick someone into answering