衡量搜索相关性

[How in the heck do you measure search relevance?](https://www.reddit.com/r/RedditEng/comments/te0gfz/how_in_the_heck_do_you_measure_search_relevance/)
[Measuring Search Relevance, Part 2: nDCG Deep Dive](https://www.reddit.com/r/RedditEng/comments/y6idrl/measuring_search_relevance_part_2_ndcg_deep_dive/)

 

We can measure the semantic similarity by calculating a cosine distance between two embedding vectors in vector space. If these vectors are close to each other then we can say that these sentences are semantically similar.

 

Improved Content Understanding and Relevance with Large Language Models (SnooBERT )

https://www.reddit.com/r/RedditEng/comments/vtjndg/improved_content_understanding_and_relevance_with/