Embedding with nested Vs flat

Plugins Machine Learning
,
1

Versions (relevant - OpenSearch/Dashboard/Server OS/Browser):
Opensearch -3.3

Describe the issue:

The embedding mapping with nested like:
“VECTOR_Nested”: {
“type”: “nested”,
“properties”: {
“embedding”: {
“type”: “knn_vector”,
“dimension”: 384,
“method”: {
“engine”: “lucene”,
“space_type”: “l2”,
“name”: “hnsw”
}
}
}
}

used with text chunking is having storage 3 times more than the Flat embedding vecctor.

Embedding Vector Flat:
“VECTOR_FLAT”: {
“type”: “knn_vector”,
“dimension”: 384,
“method”: {
“engine”: “lucene”,
“space_type”: “l2”,
“name”: “hnsw”
}
}

The chunk for the nested is only one still the storage occupied is 3 times .

Is this expected or is there any configuration I am missing.

Please let know your suggestions or feedback.

Thanks,
Tejas