[Question]: Similarity balance problem

[zone-lou]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (Language Policy).
• Non-english title submitions will be closed directly ( 非英文标题的提交将会被直接关闭 ) (Language Policy).
• Please do not modify this template :) and fill in all the required fields.

Describe your problem

When the document is parsed, the generated keywords are set. In the query, the blocks with high keyword similarity are found in the following order, and whether there is a configuration for setting keywords first.

I want the data

The actual number one data

Keyword similarity weight does not change the order of filling

[KevinHuSh]

If the keyword weight does not change the order, the embedding similarity of all the chunk must be the same.

[zone-lou]

If the keyword weight does not change the order, the embedding similarity of all the chunk must be the same.如果 keyword weight 不改变顺序，则所有 chunk 的嵌入相似度必须相同。

the embedding similarity of all the chunk must be the same. What do we do with this step