Add README

README.md

@@ -2,49 +2,95 @@
 configs:
 - config_name: amazonreviews2014_beauty
   data_files:
-  - split: validation
-    path: "data/amazonreviews2014_beauty/validation.jsonl"
+  - split: val
+    path: "AmazonReviews2014-Beauty/val.jsonl"
   - split: test
-    path: "data/amazonreviews2014_beauty/test.jsonl"
+    path: "AmazonReviews2014-Beauty/test.jsonl"
+
 - config_name: amazonreviews2014_sports_and_outdoors
   data_files:
-  - split: validation
-    path: "data/amazonreviews2014_sports_and_outdoors/validation.jsonl"
+  - split: val
+    path: "AmazonReviews2014-Sports_and_Outdoors/val.jsonl"
   - split: test
-    path: "data/amazonreviews2014_sports_and_outdoors/test.jsonl"
+    path: "AmazonReviews2014-Sports_and_Outdoors/test.jsonl"
+
 - config_name: amazonreviews2023_industrial_and_scientific
   data_files:
-  - split: validation
-    path: "data/amazonreviews2023_industrial_and_scientific/validation.jsonl"
+  - split: val
+    path: "AmazonReviews2023-Industrial_and_Scientific/val.jsonl"
   - split: test
-    path: "data/amazonreviews2023_industrial_and_scientific/test.jsonl"
+    path: "AmazonReviews2023-Industrial_and_Scientific/test.jsonl"
+
 - config_name: amazonreviews2023_musical_instruments
   data_files:
-  - split: validation
-    path: "data/amazonreviews2023_musical_instruments/validation.jsonl"
+  - split: val
+    path: "AmazonReviews2023-Musical_Instruments/val.jsonl"
   - split: test
-    path: "data/amazonreviews2023_musical_instruments/test.jsonl"
+    path: "AmazonReviews2023-Musical_Instruments/test.jsonl"
+
 - config_name: amazonreviews2023_office_products
   data_files:
-  - split: validation
-    path: "data/amazonreviews2023_office_products/validation.jsonl"
+  - split: val
+    path: "AmazonReviews2023-Office_Products/val.jsonl"
   - split: test
-    path: "data/amazonreviews2023_office_products/test.jsonl"
+    path: "AmazonReviews2023-Office_Products/test.jsonl"
+
 - config_name: steam
   data_files:
-  - split: validation
-    path: "data/steam/validation.jsonl"
+  - split: val
+    path: "Steam/val.jsonl"
   - split: test
-    path: "data/steam/test.jsonl"
-- config_name: yelp_yelp_2020
+    path: "Steam/test.jsonl"
+
+- config_name: yelp_2020
   data_files:
-  - split: validation
-    path: "data/yelp_yelp_2020/validation.jsonl"
+  - split: val
+    path: "Yelp-Yelp_2020/val.jsonl"
   - split: test
-    path: "data/yelp_yelp_2020/test.jsonl"
+    path: "Yelp-Yelp_2020/test.jsonl"
 ---
 
-# MemGen annotation datasets
+# MemGen Annotations
+
+This is the annotation dataset for **On the Memorization and Generalization of Generative Recommendation**.
+
+\[[Paper]\] \[[Code]\] \[[Models]\]
+
+[Paper]: <link not ready yet>
+[Code]: https://github.com/Jamesding000/MemGen-GR
+[Models]: https://huggingface.co/jamesding0302/memgen-checkpoints
+
+The annotations categorize evaluation instances under the leave-one-out protocol:
+- **test** split uses the **last** item in the user history sequence as target,
+- **val** split uses the **second-to-last** item as target.
+
+## Columns
+- `sample_id`: row index within the split in the original dataset.
+- `user_id`: raw user identifier (join key).
+- `master`: one of `memorization`, `generalization`, `uncategorized`.
+- `subcategories`: list of `{rule, hop}` for fine-grained generalization types.
+- `all_labels`: all string labels (e.g., `["generalization", "symmetry_3"]`).
+
+## Load
+```python
+from datasets import load_dataset
+
+labels = load_dataset(
+    "jamesding0302/memgen-annotations",
+    "amazonreviews2014_beauty",
+    split="test",
+)
+print(labels[0])
+```
+
+## Merge with processed dataset
+```python
+from datasets import load_dataset
+
+lab = load_dataset("json", data_files="item_case_records.jsonl", split="train", cache_dir="./hf_cache")
 
-This repo contains memorization/generalization annotations for multiple benchmarks.
-Each benchmark is a config (subset) selectable in the HF dataset dropdown.
+ds = pipeline.split_datasets["train"]
+ds = ds.add_column("master", lab["master"]) \
+       .add_column("subcategories", lab["subcategories"]) \
+       .add_column("all_labels", lab["all_labels"])
+```