Sparse vector search scoredPoint scores

[nirmalavhcl]

I have a collection with dense and sparse vector embeddings. For a given query, I generate the dense and sparse vector and search against the corresponding dense and sparse vectors in the collection using queryAsync call(Java client). With my Natural language query, typically there are 4-5 values in the sparse vector. I notice that the scores of the sparse vectors are way low( in the 0.5 range) as compared to the one against dense vector search(above 0.8 range). With my query, I only see 2 matches or so that intersect between the two searches. Is the scoring values for sparse vectors in that low range normal?