Update README.md

README.md

@@ -1150,7 +1150,7 @@ English, French, Italian, German, Portuguese, Spanish
 
 ## Dataset Structure
 
-It is structured in the following format: {language}_{type}_{shard}.jsonl.xz
+It is structured in the following format: {language}\_{type}\_{shard}.jsonl.xz
 
 type is one of the following:
 - laws
@@ -1159,7 +1159,7 @@ type is one of the following:
 Use the the dataset like this:
 ```
 from datasets import load_dataset
-config = 'fr_laws' #{language}_{type}
+config = 'fr_laws' #{language}\_{type}
 dataset = load_dataset('rdcs/MultiLegalSBD', config)
 ```
 ### Data Instances
@@ -1168,7 +1168,20 @@ dataset = load_dataset('rdcs/MultiLegalSBD', config)
 
 ### Data Fields
 
-[More Information Needed]
+- text: the original text
+- spans:
+  - start: offset of the first character
+  - end: offset of the last character
+  - label: "Sentence"
+  - token_start: id of the first token
+  - token_end: id of the last token
+- tokens:
+  - text: token text
+  - start: offset of the first character
+  - end: offset of the last character
+  - id: token id
+  - ws: whether the token is followed by whitespace
+  
 
 ### Data Splits