Update README.md

README.md

@@ -65,7 +65,7 @@ This dataset can be used directly with Sentence Transformers to train embedding
 ### Split Details
 
 * `train` split: 3,007,396 pairs
-* `eval` split: 2,500 pairs
-* `test` split: 2,500 pairs
+* `eval` split: 2,500 pairs (all unique)
+* `test` split: 2,500 pairs (all unique)
 
 There are 0 common questions across the `train`, `eval`, and `test` splits.