fix readme

README.md

@@ -6,14 +6,14 @@ license:
 multilinguality:
 - monolingual
 size_categories:
-- 1K<n<10K
+- n<1K
 pretty_name: ConceptNet with High Confidence
 ---
 # Dataset Card for "relbert/conceptnet_relation_similarity"
 ## Dataset Description
 - **Repository:** [RelBERT](https://github.com/asahi417/relbert)
 - **Paper:** [https://home.ttic.edu/~kgimpel/commonsense.html](https://home.ttic.edu/~kgimpel/commonsense.html)
-- **Dataset:** High Confidence Subset of ConceptNet
+- **Dataset:** Relational similarity dataset based on the high-confidence subset of ConceptNet
 
 ### Dataset Summary
 The selected subset of ConceptNet used in [this work](https://home.ttic.edu/~kgimpel/commonsense.html), which compiled 
@@ -21,15 +21,11 @@ to fine-tune [RelBERT](https://github.com/asahi417/relbert) model.
 We removed `NotCapableOf` and `NotDesires` to keep the positive relation only.
 We consider the original test set as test set, dev1 as the training set, and dev2 as the validation set.
 
-|train|validation|test|
-|--------:|----:|---------:|
-|15| 17 |      15|
-
 
 ## Dataset Structure
 ### Data Instances
 An example of `train` looks as follows.
-```
+```shell
 {
     "relation_type": "AtLocation",
     "positives": [["fish", "water"], ["cloud", "sky"], ["child", "school"], ... ],
@@ -37,6 +33,11 @@ An example of `train` looks as follows.
 }
 ```
 
+### Data Splits
+
+|   train |validation|      test|
+|--------:|---------:|---------:|
+|       15| 17       |        15|
 
 
 ### Citation Information