Dataset Featurization Added

README.md

@@ -0,0 +1,100 @@
+pretty_name: Dataset Featurization
+language:
+  - en
+license:
+  - mit
+task_categories:
+  - feature-extraction
+task_ids:
+  - language-modeling
+configs:
+  - config_name: nyt
+    data_files:
+      - split: train
+        path: data/nyt/samples.csv
+  - config_name: nyt-evaluation-0
+    data_files:
+      - split: train
+        path: data/evaluation/evaluation_df_group_0.csv
+  - config_name: nyt-evaluation-1
+    data_files:
+      - split: train
+        path: data/evaluation/evaluation_df_group_1.csv
+  - config_name: nyt-evaluation-2
+    data_files:
+      - split: train
+        path: data/evaluation/evaluation_df_group_2.csv
+  - config_name: amazon
+    data_files:
+      - split: train
+        path: data/amazon/samples.csv
+  - config_name: amazon-evaluation-0
+    data_files:
+      - split: train
+        path: data/evaluation/evaluation_df_group_0.csv
+  - config_name: amazon-evaluation-1
+    data_files:
+      - split: train
+        path: data/evaluation/evaluation_df_group_1.csv
+  - config_name: amazon-evaluation-2
+    data_files:
+      - split: train
+        path: data/evaluation/evaluation_df_group_2.csv
+  - config_name: dbpedia
+    data_files:
+      - split: train
+        path: data/dbpedia/samples.csv
+  - config_name: dbpedia-evaluation-0
+    data_files:
+      - split: train
+        path: data/evaluation/evaluation_df_group_0.csv
+  - config_name: dbpedia-evaluation-1
+    data_files:
+      - split: train
+        path: data/evaluation/evaluation_df_group_1.csv
+  - config_name: dbpedia-evaluation-2
+    data_files:
+      - split: train
+        path: data/evaluation/evaluation_df_group_2.csv
+    
+# Dataset Featurization: Experiments
+
+This repository contains datasets used in evaluating **Dataset Featurization** against the prompting baseline. For datasets used in the case studies, please refer to [Compositional Preference Modeling](https://huggingface.co/datasets/Bravansky/compositional-preference-modeling) and [Compact Jailbreaks](https://huggingface.co/datasets/Bravansky/compact-jailbreaks).
+
+The evaluation focuses on three datasets: The [New York Times Annotated Corpus (NYT)](https://catalog.ldc.upenn.edu/docs/LDC2008T19/new_york_times_annotated_corpus.pdf), [Amazon Reviews (Amazon)](https://amazon-reviews-2023.github.io/), and [DBPEDIA](https://huggingface.co/datasets/DeveloperOats/DBPedia_Classes). For each dataset, we sample 15 different categories and construct three separate subsets, each containing 5 categories with 1000 samples per category. We evaluate the featurization method's performance on each subset.
+
+### NYT
+
+From the NYT corpus, we utilize manually reviewed tags from the NYT taxonomy classifier, specifically focusing on articles under "Features" and "News" categories, to construct a dataset of texts with their assigned categories. Below is how to access the input datasets and the proposed features with their assignments from the evaluation stage:
+
+```python
+import datasets
+text_df = load_dataset("Bravansky/compositional-preference-modeling", "nyt", split="train").to_pandas()
+evaluation_df_0 = load_dataset("Bravansky/compositional-preference-modeling", "nyt-evaluation-0", split="train").to_pandas()
+evaluation_df_1 = load_dataset("Bravansky/compositional-preference-modeling", "nyt-evaluation-1", split="train").to_pandas()
+evaluation_df_2 = load_dataset("Bravansky/compositional-preference-modeling", "nyt-evaluation-2", split="train").to_pandas()
+```
+
+### Amazon
+
+Using a dataset of half a million customer reviews, we focus on identifying high-level item categories (e.g., Books, Fashion, Beauty), excluding reviews labeled "Unknown". The input datasets and the proposed features with their assignments from the evaluation stage can be accessed as follows:
+
+```python
+import datasets
+text_df = load_dataset("Bravansky/compositional-preference-modeling", "amazon", split="train").to_pandas()
+evaluation_df_0 = load_dataset("Bravansky/compositional-preference-modeling", "amazon-evaluation-0", split="train").to_pandas()
+evaluation_df_1 = load_dataset("Bravansky/compositional-preference-modeling", "amazon-evaluation-1", split="train").to_pandas()
+evaluation_df_2 = load_dataset("Bravansky/compositional-preference-modeling", "amazon-evaluation-2", split="train").to_pandas()
+```
+
+### DBPEDIA
+
+Using the pre-processed DBPEDIA dataset, we focus on reconstructing categories labeled as level `l2`:
+
+```python
+import datasets
+text_df = load_dataset("Bravansky/compositional-preference-modeling", "dbpedia", split="train").to_pandas()
+evaluation_df_0 = load_dataset("Bravansky/compositional-preference-modeling", "dbpedia-evaluation-0", split="train").to_pandas()
+evaluation_df_1 = load_dataset("Bravansky/compositional-preference-modeling", "dbpedia-evaluation-1", split="train").to_pandas()
+evaluation_df_2 = load_dataset("Bravansky/compositional-preference-modeling", "dbpedia-evaluation-2", split="train").to_pandas()
+```