Upload README.md with huggingface_hub

README.md

@@ -1,35 +1,50 @@
 ---
-dataset_info:
-  features:
-  - name: chosen_timestamp
-    dtype: timestamp[ns]
-  - name: rejected_timestamp
-    dtype: timestamp[ns]
-  - name: score_chosen
-    dtype: int64
-  - name: score_rejected
-    dtype: int64
-  - name: chosen
-    list:
-    - name: content
-      dtype: string
-    - name: role
-      dtype: string
-  - name: rejected
-    list:
-    - name: content
-      dtype: string
-    - name: role
-      dtype: string
-  splits:
-  - name: train
-    num_bytes: 288813335
-    num_examples: 43827
-  download_size: 181309703
-  dataset_size: 288813335
+pretty_name: LitBench Preference Pairs (chat format)
+dataset_name: CarlOwOs/LitBench-Processed
+task_categories:
+  - text-generation
+  - preference-modeling
+tags:
+  - rlhf
+  - pairwise-preference
+  - conversations
+  - literature
+language:
+  - en
+size_categories:
+  - n<1M
 configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
+  - config_name: default
+    data_files: auto
 ---
+
+# LitBench Preference Pairs (chat format)
+
+This dataset consolidates the LitBench train and test preference pairs into one repository with two splits: `train` and `test`. Each example contains two chat-style conversations under keys `chosen` and `rejected`, plus numeric scores `score_chosen` and `score_rejected`.
+
+Last updated: 2025-10-04 05:10 UTC
+
+## Splits
+
+- train: 43827 examples
+- test: 0 examples
+
+## Features
+
+- `chosen`: list of role/content messages representing the preferred conversation
+- `rejected`: list of role/content messages representing the alternative conversation
+- `score_chosen`: integer upvote score of the chosen story
+- `score_rejected`: integer upvote score of the rejected story
+
+## Loading
+
+```python
+from datasets import load_dataset
+
+ds = load_dataset("CarlOwOs/LitBench-Processed")
+print(ds)
+```
+
+## Citation
+
+If you use this dataset, please also acknowledge the original LitBench sources.