The community member is using VectorStoreIndex to evaluate an embedding model and is curious about the default similarity function. They noticed that the node scores are all between 0 and 1, rather than the expected range of -1 to 1 for cosine similarity. The community members suggest that the similarity function may be using a normalized L2 function instead. They also provide information on how to change the default similarity function, by subclassing the embedding model and changing the similarity mode. However, the original community member still has a question about why the similarity scores are limited to the 0-1 range, even when using the default cosine similarity function.
Hello, I use VectorStoreIndex to evaluate embedding model. I think default similarity function is cos similarity. But as I check node score, they are all between (0, 1) rather than (-1, 1). It seems like maybe normalized L2 function is being used. Why all similarity score is between 0-1? Can I change default similarity function of VectorStoreIndex to some specific function instead of default cos function?
Thanks a lot. With your kind help now I know how to change similarity fn. But still 1 more question. Why all similarity scores are between (0, 1). Since I am using default COSINE fn, it should be in (-1, 1) right?
