Add dataset summary and field tables

README.md

@@ -92,6 +92,17 @@ OmniToM should not be treated as:
 ## Dataset Structure
 
 
+### Overall Statistics
+
+| Measure | Value |
+| --- | ---: |
+| Released stories | 895 |
+| Story categories | 7 |
+| Annotated belief propositions | 22,343 |
+| Schema dimensions per proposition | 7 |
+| Total schema labels | 156,401 |
+
+
 ### Data Instances
 
 
@@ -124,34 +135,21 @@ Each line in the release file is one JSON object:
 
 ### Data Fields
 
-
-- `story_id`
-  - Unique integer identifier for the story.
-- `story_category`
-  - One of seven retained benchmark categories.
-- `story`
-  - Raw story text used for both extraction and labeling tasks.
-- `beliefs`
-  - List of annotated belief propositions.
-- `beliefs[].actor`
-  - Belief holder. The reserved actor `world` denotes narrator/world facts.
-- `beliefs[].belief`
-  - Minimal propositional belief statement.
-- `beliefs[].labels.order`
-  - Recursive depth in `{0,1,2,3}`.
-- `beliefs[].labels.truth_status`
-  - `True`, `False`, or `Unknown`.
-- `beliefs[].labels.knowledge_access`
-  - `Private`, `Shared`, or `Public`.
-- `beliefs[].labels.representation`
-  - `Explicit` or `Implicit`.
-- `beliefs[].labels.content_type`
-  - One of: `Location`, `Contents/Physical State`, `Identity/Relation`, `Epistemic`, `Desire/Intention`, `Emotion`, `Trait/Value`, `Action/Event`.
-- `beliefs[].labels.mental_source`
-  - One of: `Narration`, `Perception`, `Memory`, `Testimony`, `Inference`, `Imagination`, `Unknown`.
-- `beliefs[].labels.context`
-  - `Neutral`, `Temporal`, `Deceptive`, or `Counterfactual`.
-
+| Field | Type | Description or allowed values |
+| --- | --- | --- |
+| `story_id` | Integer | Unique story identifier. |
+| `story_category` | String | One of the seven story categories listed below. |
+| `story` | String | Story text used for belief extraction and labeling. |
+| `beliefs` | List of objects | Annotated belief propositions for the story. |
+| `beliefs[].actor` | String | Belief holder. `world` marks narrator or world facts. |
+| `beliefs[].belief` | String | Minimal propositional belief statement. |
+| `beliefs[].labels.order` | String | Recursive depth: `0`, `1`, `2`, or `3`. |
+| `beliefs[].labels.truth_status` | String | `True`, `False`, or `Unknown`. |
+| `beliefs[].labels.knowledge_access` | String | `Private`, `Shared`, or `Public`. |
+| `beliefs[].labels.representation` | String | `Explicit` or `Implicit`. |
+| `beliefs[].labels.content_type` | String | `Location`, `Contents/Physical State`, `Identity/Relation`, `Epistemic`, `Desire/Intention`, `Emotion`, `Trait/Value`, or `Action/Event`. |
+| `beliefs[].labels.mental_source` | String | `Narration`, `Perception`, `Memory`, `Testimony`, `Inference`, `Imagination`, or `Unknown`. |
+| `beliefs[].labels.context` | String | `Neutral`, `Temporal`, `Deceptive`, or `Counterfactual`. |
 
 ### Data Splits
 
@@ -180,17 +178,18 @@ OmniToM instead evaluates explicit belief-structure modeling. The benchmark is g
 ### Source Data
 
 
-The benchmark sources its stories from ToMBench. From the original source corpus, OmniToM retains seven story categories whose stories provide sufficiently self-contained mental-state evidence for belief extraction from text alone:
-
-
-- Ambiguous Story Task
-- False Belief Task
-- Faux-pas Recognition Test
-- Hinting Task Test
-- Persuasion Story Task
-- Scalar Implicature Test
-- Strange Story Task
+The benchmark sources its stories from ToMBench. It retains seven categories whose stories provide sufficiently self-contained mental-state evidence for belief extraction from text alone.
 
+| Story category | Stories |
+| --- | ---: |
+| Ambiguous Story Task | 98 |
+| False Belief Task | 97 |
+| Faux-pas Recognition Test | 142 |
+| Hinting Task Test | 100 |
+| Persuasion Story Task | 97 |
+| Scalar Implicature Test | 154 |
+| Strange Story Task | 207 |
+| **Total** | **895** |
 
 ### Annotation Process
 
@@ -274,3 +273,5 @@ Code and evaluation scripts: [BridgeAI-Lab/OMNITOM](https://github.com/BridgeAI-
 
 Project page: [OmniToM](https://adam-12-0.github.io/omnitom-project/)
 
+Camera-ready Croissant metadata with Responsible AI fields: [Download JSON](https://huggingface.co/datasets/BridgeAI-Lab/omnitom-benchmark/resolve/main/croissant_rai_omnitom_camera_ready.json)
+