Create mimir.py

mimir.py

@@ -0,0 +1,98 @@
+"""
+    Data used for experiments with MIMIR. Processed train/test splits for models trained on the Pile (for now).
+    Processing data at HF end.
+"""
+
+from datasets import GeneratorBasedBuilder, SplitGenerator, DownloadManager, BuilderConfig
+import json
+import os
+
+import datasets
+
+
+_HOMEPAGE = "http://github.com/iamgroot42/mimir"
+
+_DESCRIPTION = """\
+Member and non-member splits for our MI experiments using MIMIR. Data is available for each source.
+We also cache neighbors (generated for the NE attack).
+"""
+
+_CITATION = """\
+@article{duan2024do,
+  title={Do Membership Inference Attacks Work on Large Language Models?},
+  author={Duan*, Michael and \textbf{A. Suri*} and Mireshghallah, Niloofar and Min, Sewon and Shi, Weijia and Zettlemoyer, Luke and Tsvetkov, Yulia and Choi, Yejin and Evans, David and Hajishirzi, Hannaneh},
+  journal={arXiv preprint arXiv:???},
+  year={2024}
+}
+"""
+
+class MimirConfig(BuilderConfig):
+    """BuilderConfig for Mimir dataset."""
+
+    def __init__(self, **kwargs):
+        """Constructs a MimirConfig.
+
+        Args:
+            **kwargs: keyword arguments forwarded to super.
+        """
+        super(MimirConfig, self).__init__(**kwargs)
+
+
+class MimirDataset(GeneratorBasedBuilder):
+    # Assuming 'VERSION' is defined
+    VERSION = datasets.Version("1.0.0")
+
+    # Define the builder configs
+    BUILDER_CONFIG_CLASS = MimirConfig
+    BUILDER_CONFIGS = [
+        MimirConfig(name="the_pile_arxiv", description="This split contains data from Arxiv, truncated with 7-gram overlap threshold < 0.2."),
+    ]
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            # This is the description that will appear on the datasets page.
+            description=_DESCRIPTION,
+            # This defines the different columns of the dataset and their types
+            features=datasets.Features(
+                {
+                    "text": datasets.Value("string"),  # Each example is a piece of text
+                }
+            ),
+            # If there's a common (input, target) tuple from the features,
+            # specify them here. They'll be used if as_supervised=True in
+            # builder.as_dataset.
+            supervised_keys=None,
+            # Homepage of the dataset for documentation
+            homepage=_HOMEPAGE,
+            # Citation for the dataset
+            # citation=_CITATION,
+        )
+
+    def _split_generators(self, dl_manager: DownloadManager):
+        """Returns SplitGenerators."""
+        # Path to the data files
+        NEIGHBOR_SUFFIX = "_neighbors_25_bert_in_place_swap"
+        parent_dir = "cache_100_200_1000_512"
+
+        file_paths = {
+            "member": os.path.join(parent_dir, "train", self.config.name + ".jsonl"),
+            "nonmember": os.path.join(parent_dir, "test", self.config.name + ".jsonl")
+        }
+        # Load neighbor splits if they exist
+        if os.path.exists(os.path.join(parent_dir, "train_neighbors", self.config.name + f"{NEIGHBOR_SUFFIX}.jsonl")):
+            # Assume if train nieghbors exist, test neighbors also exist
+            file_paths["member_neighbors"] = os.path.join("cache_100_200_1000_512", "train_neighbors", self.config.name + f"{NEIGHBOR_SUFFIX}.jsonl"),
+            file_paths["nonmember_neighbors"] = os.path.join("cache_100_200_1000_512", "test_neighbors", self.config.name + f"{NEIGHBOR_SUFFIX}.jsonl")
+
+        splits = []
+        for k, v in file_paths.items():
+            splits.append(SplitGenerator(name=k, gen_kwargs={"file_path": v}))
+        return splits
+
+    def _generate_examples(self, file_path):
+        """Yields examples."""
+        # Open the specified .jsonl file and read each line
+        with open(file_path, 'r') as f:
+            for id, line in enumerate(f):
+                data = json.loads(line)
+                yield id, {"text": data}