diff --git a/NLP_documentation.md b/NLP_documentation.md
index 42cc6ca2e01aff1a192c5a2d82ffe84cf71f3143..b8cf1196dd9a2995c6bfaf93d84e92f72159c456 100644
--- a/NLP_documentation.md
+++ b/NLP_documentation.md
@@ -202,7 +202,7 @@ The average number of overlaps (full, min and max) and the average weighted simi
 | `pairwise_max` | ![pairwise_max](/images/pairwise_max_heatmap.png) |
 | `pairwise_weighted` | ![pairwise_weighted](/images/pairwise_weighted_heatmap.png) |
 
-For word-level analysis four additional dataframes are created for one of the six model combinations, namely LatinISE for a general language model and ROMTEXT for a legal language model. The words in the dataframes are sorted in descending order of overlap or weighted similarity. Those on the top are the words which feature prominently in both models.
+For word-level analysis four additional dataframes are created for one of the six model combinations, namely the combination of LatinISE and ROMTEXT, one general and one legal language model. The words in the dataframes are sorted in descending order of overlap or weighted similarity. Those on the top are the words which feature prominently in both models.
 
 ### Footnotes