is it possible you can share an example of extracting general text embedding ? I would love to test the clustering task based on semantic similarity