Setting Score Thresholds for Text Segments in Vector Matching: Still Encountering Redundant Parts

[20246688]

Blogger, the vector matching similarity will assign a score to each text segment, so that all text segments have similarity values. May I ask how I can feed the text segments I need to the model? Should I set a score threshold? But there will still be some redundant parts.

[mnorris11]

You may get more help from Blogger. We don't have context on feeding text segments to models, just on the KNN vector search.

[20246688]

您可能会从 Blogger 获得更多帮助。我们没有关于将文本段提供给模型的上下文，只有关于 KNN 向量搜索。

thanks