Update README.md

README.md

@@ -88,3 +88,26 @@ Each instance in the dataset contains the following primary fields:
     *   `task_id` or `question_id_original`: Original identifier for the question/task.
     *   `correct_answer`: The correct option label, if applicable (e.g., for problem-solving tasks).
     *   Other dataset-specific information.
+
+
+## How to Use
+
+You can load SimBench using the `datasets` library. Since the data is stored in `.pkl` files, ensure you have a local copy of the dataset loading script (typically `SimBench.py` or the name of your dataset on the Hub) in your working directory or Python path if `datasets` cannot automatically find it for this format.
+
+```python
+from datasets import load_dataset
+
+# Load the 'default' configuration which contains both SimBenchPop and SimBenchGrouped
+simbench_data_dict = load_dataset("pitehu/SimBench", name="default")
+
+# Access SimBenchPop
+simbench_pop = simbench_data_dict["SimBenchPop"]
+
+# Access SimBenchGrouped
+simbench_grouped = simbench_data_dict["SimBenchGrouped"]
+
+# Example: Accessing the first instance in SimBenchPop
+instance = simbench_pop[0]
+print(instance)
+
+