Cluster Name Change in exported csv

I am doing some clustering analysis using the Weka-EM operator. I am joining my original data to the cluster field attributed to each row and exoprting this data using the 'write-csv' operator.  I noticed that the cluster names (e.g. cluster 0, cluster 1 ...) are different in the example data in RM to that in the exported CSV. The clusters are obviously the same but the names attributed are different!  ???