Update README.md

README.md

@@ -33,4 +33,26 @@ Each data file ({train/dev/test}.jsonl) is formatted as a JSON lines file. Each
 questions: a list of questions and accompanying responses
   - question text
   - question number: the order in which that question was answered by the writers
-  - responses: list of worker's response, where each response is a dictionary containing the (anonymized) worker ID, an internal UID, and their response to the question
+  - responses: list of worker's response, where each response is a dictionary containing the (anonymized) worker ID, an internal UID, and their response to the question
+ 
+
+### dataset contents
+
+
+```python
+DatasetDict({
+    train: Dataset({
+        features: ['metadata', 'document', 'questions'],
+        num_rows: 50
+    })
+    test: Dataset({
+        features: ['metadata', 'document', 'questions'],
+        num_rows: 52
+    })
+    validation: Dataset({
+        features: ['metadata', 'document', 'questions'],
+        num_rows: 25
+    })
+})
+```
+