Is any solutions for fixing such incorrect behaviour on semantic vector search.
For example. thenlper/gte-large
model = SentenceTransformer(‘thenlper/gte-large’)
query: “two point eight”
results
[ScoredPoint(id=1, version=0, score=0.9332634, ‘vector_data’: ‘eight point two patch, one .’}, vector=None),
ScoredPoint(id=2, version=0, score=0.92371386, ‘vector_data’: ‘two point eight’}, vector=None),
ScoredPoint(id=0, version=0, score=0.9218446, ‘vector_data’: 'eight point two patch, zero '}, vector=None),