First version of 2D-ATOMS dataset

.gitattributes

@@ -53,3 +53,13 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.jpg filter=lfs diff=lfs merge=lfs -text
 *.jpeg filter=lfs diff=lfs merge=lfs -text
 *.webp filter=lfs diff=lfs merge=lfs -text
+task1_short_term_intention.json filter=lfs diff=lfs merge=lfs -text
+task2_long_term_intention.json filter=lfs diff=lfs merge=lfs -text
+task4_perception.json filter=lfs diff=lfs merge=lfs -text
+task5_first_order_belief.json filter=lfs diff=lfs merge=lfs -text
+task9_emotions.json filter=lfs diff=lfs merge=lfs -text
+task0_action_understanding.json filter=lfs diff=lfs merge=lfs -text
+task3_desire.json filter=lfs diff=lfs merge=lfs -text
+task6_second_order_belief.json filter=lfs diff=lfs merge=lfs -text
+task7_non_literal_communication.json filter=lfs diff=lfs merge=lfs -text
+task8_knowledge.json filter=lfs diff=lfs merge=lfs -text

2D-ATOMS.py

@@ -0,0 +1,55 @@
+import datasets
+
+_DESCRIPTION = """\
+We introduce **2D-ATOMS** dataset, a novel text-based dataset that evaluates machine's reasoning process under situated theory-of-mind setting.
+
+Our dataset includes 9 different ToM evaluation tasks for each mental state under ATOMS[1], and 1 reality-checking task to test LLMs’ understanding of the world. It is important to acknowledge that our experiment serves as a proof of concept and does not aim to cover the entire spectrum of machine ToM, as our case studies are far from being exhaustive or systematic. Here we release the zero-shot version of our dataset, which is used in our paper.
+"""
+
+class TRIP(datasets.GeneratorBasedBuilder):
+    VERSION = datasets.Version("1.0.1")
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=datasets.Features(
+                {
+                    "prompt": datasets.Value("string"),
+                    "answer": datasets.Value("string"),
+                      
+                }
+            ),
+        )
+    
+    def _split_generators(self, dl_manager: datasets.DownloadManager):
+        """Returns SplitGenerators."""
+        task_list = [
+            "task0_action_understanding",
+            "task1_short_term_intention",
+            "task2_long_term_intention",
+            "task3_desire",
+            "task4_perception",
+            "task5_first_order_belief",
+            "task6_second_order_belief",
+            "task7_non_literal_communication",
+            "task8_knowledge",
+            "task9_emotions"
+        ]
+        data_roots = dl_manager.download_and_extract({k: f"{k}.json" for k in task_list})
+        return [
+            datasets.SplitGenerator(
+                name=split,
+                gen_kwargs={
+                    "filepath": data_roots[split],
+                },
+            )
+            for split in task_list
+        ]
+
+    def _generate_examples(self, filepath):
+        # load jsonl file
+        import json
+        with open(filepath) as f:
+            data = [json.loads(line) for line in f]
+        for i, example in enumerate(data):
+            yield i, example

README.md

@@ -0,0 +1,23 @@
+# 2D-ATOMS: 2D Abilities in Theory of Mind Space dataset
+
+Official dataset for **Towards A Holistic Landscape of Situated Theory of Mind in Large Language Models**. Ziqiao Ma, Jacob Sandom, Run Peng, Joyce Chai. EMNLP Findings, 2023.
+
+## Overview
+
+![image](overview_hf.png)
+
+We introduce **2D-ATOMS** dataset, a novel text-based dataset that evaluates machine's reasoning process under situated theory-of-mind setting.
+
+Our dataset includes 9 different ToM evaluation tasks for each mental state under ATOMS[1], and 1 reality-checking task to test LLMs’ understanding of the world. It is important to acknowledge that our experiment serves as a proof of concept and does not aim to cover the entire spectrum of machine ToM, as our case studies are far from being exhaustive or systematic. Here we release the zero-shot version of our dataset, which is used in our paper.
+
+## Download
+
+```python
+from datasets import load_dataset
+dataset = load_dataset("sled-umich/2D-ATOMS")
+```
+
+## Reference
+
+
+[1] C. Beaudoin, É. Leblanc, C. Gagner, and M. H. Beauchamp, ‘Systematic review and inventory of theory of mind measures for young children’, Frontiers in psychology, vol. 10, p. 2905, 2020.

task0_action_understanding.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e6a769da98e9550e0e06e6acf32a203a076d38451f70aee6600539f73cfd27c
+size 288435

task1_short_term_intention.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f615059e060847ec88dcdb7eeedad9b80932f844f9063d5a28cf063a78687dd5
+size 229290

task2_long_term_intention.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c04c64622de4c8b0e7614d83b643072da401d8e19433e5116011ac49bfe3ee9
+size 337843

task3_desire.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ecb982b1e1b09af876848211d1c4afd97df22e48927649d36db48b39a1ae9e5
+size 358127

task4_perception.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fad90f699cf818fd364dfa0245699e17751731b1e2e3cbebb1406f0150afced3
+size 414223

task5_first_order_belief.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba670d0599ddaece4f1d0964c81dba64e008b32ff66b3ca6a7e4bd4df93ca07d
+size 881046

task6_second_order_belief.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b6fb81ec9d2e419a31c8015feb581cd9c7395cba1a48f92b773c6e4c51e4f94
+size 1343110

task7_non_literal_communication.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b50d7ce14c2ae327d5fe987830acb93cb8bb404f4e198e690f79d90098cdcf75
+size 309783

task8_knowledge.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:327e35d220808de3e4dd19ae2a75d6e45ee677ce8131f23d7cd1d13e38bf18af
+size 571240

task9_emotions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2eff77d07a558629513e0b8840cc6331f6f38140fbf467dc22052930982e1d1d
+size 745243