"NA" matches "naive CD4+ T cell" in CL

[cmungall]

$ curl -L -s 'http://www.ebi.ac.uk/spot/zooma/v2/api/services/annotate?propertyValue=NA' | jq '.[] | .confidence, .semanticTags, .annotatedProperty.propertyValue'
"GOOD"
[
  "http://purl.obolibrary.org/obo/CL_0000895"
]
"na??ve CD4+ T cell"

I think it is treating the umlauted i as a token separator and doing a substring match... but even with this explanation, why is this GOOD?

[cmungall]

actually I am not sure my explanation is correct

https://www.ebi.ac.uk/ols/ontologies/cl/terms?iri=http%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FCL_0000895

naive CD4+ T cell is a broad synonym

still unclear why this is a GOOD match