知识卡片

普通读书笔记卡 · 8658.a · 生成模型:GPT-6

扰动输入只能支持限定诊断

理解等级:中级

添加、删除或替换token后比较输出,可发现输入敏感性和提示问题。但扰动可能改变语义、长度或分布,不能自动解释为现实事件的因果效应;应一次控制可解释因素并保留原样本。

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

      we would like to be able to understand model behavior and explore the examples for which the model fails. Interpretability in LLMs is much less developed than in other areas of machine learning. However, we can get partial interpretability by exploring how the output changes upon minor variances in the input, and by analyzing the intermediate outputs as the inputs propagate through the Transformer architecture.

    参考来源

    - 位置:《Designing Large Language Model Applications》第05章《按任务适配模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch05.html) - 结论依据:摘录中关于扰动输入只能支持限定诊断的定义、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:we would like to be able to understand model behavior and explore the examples for which the model fails. Interpretability in LLMs is much less developed than in other areas of machine learning. However, we can get partial interpretability by exploring how the output changes upon minor variances in the input, and by analyzing the intermediate outputs as the inputs propagate through the Transformer architecture.