← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8699.d · 生成模型:GPT-6

语义相近不保证否定关系一致

理解等级:中级

同一人物和主题可能使相反陈述仍有高相似度。若任务依赖是否退休、是否允许等关键关系,应专门评估否定和矛盾样本;增加top-k只是扩大候选,不保证模型选对或矛盾被解决。

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

      After his 25th anniversary at the company, Mr. Pomorenko confirmed that he is not retiring. Mr. Pomorenko announced his retirement yesterday. Mr. Pomorenko did not announce his retirement yesterday. Now let’s use the Sentence Transformers all-mpnet-base-v2 embedding model to encode these sentences and calculate their similarity: !pip install sentence-transformers from sentence_transformers import SentenceTransformer, util model = SentenceTransformer('all-mpnet-base-v2') sentences = ['After his 25th anniversary at the

    参考来源

    - 位置:《Designing Large Language Model Applications》第11章《表示学习与嵌入》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch11.html) - 结论依据:摘录中关于语义相近不保证否定关系一致的定义、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:After his 25th anniversary at the company, Mr. Pomorenko confirmed that he is not retiring. Mr. Pomorenko announced his retirement yesterday. Mr. Pomorenko did not announce his retirement yesterday. Now let’s use the Sentence Transformers all-mpnet-base-v2 embedding model to encode these sentences and calculate their similarity: !pip install sentence-transformers from sentence_transformers import SentenceTransformer, util model = SentenceTransformer('all-mpnet-base-v2') sentences = ['After his 25th anniversary at the