Update README.md

README.md

@@ -12,20 +12,31 @@ pretty_name: RottenReviews
 size_categories:
 - 10K<n<100K
 configs:
-- config_name: OpenReview
+- config_name: ICLR2024
   data_files:
-  - split: ICLR2024
+  - split: data
     path:
     - raw/iclr2024_submissions.jsonl
-  - split: NIPS2023
+- config_name: NIPS2023
+  data_files:
+  - split: data
     path:
     - raw/neurips2023_submissions.jsonl
-- config_name: F1000
-  data_files: raw/f1000research_submissions.jsonl
-- config_name: SWJ
-  data_files: raw/semantic-web-journal_submissions.jsonl
-- config_name: HA
-  data_files: human_annotation_data.jsonl
+- config_name: F1000Journal
+  data_files:
+  - split: data
+    path:
+    - raw/f1000research_submissions.jsonl
+- config_name: SemanticWebJournal
+  data_files:
+  - split: data
+    path:
+    - raw/semantic-web-journal_submissions.jsonl
+- config_name: human_annotated_data
+  data_files: 
+  - split: data
+    path:
+    - human_annotation_data.jsonl
 ---
 
 # RottenReviews: Benchmarking Review Quality with Human and LLM-Based Judgments
@@ -56,15 +67,11 @@ The dataset was introduced to support research on benchmarking and modeling peer
 ```python
 from datasets import load_dataset
 
-dataset = load_dataset("Reviewerly/RottenReviews")
+dataset = load_dataset("Reviewerly/RottenReviews", "ICLR2024") # Select partiotion from ['ICLR2024', 'NIPS2023', 'F1000Journal', 'SemanticWebJournal', 'human_annotated_data']
 
 # Access processed reviews
-processed_reviews = dataset["processed"]
+processed_reviews = dataset["data"]
 print(processed_reviews[0])
-
-# Access human annotations
-human_data = dataset["human_annotation_data"]
-print(human_data[0])
 ```