Upload README.md with huggingface_hub

README.md

@@ -1,27 +1,90 @@
 ---
 dataset_info:
   features:
-  - name: id
-    dtype: string
-  - name: type
-    dtype: string
-  - name: name_nodes
-    dtype: string
-  - name: mean_nodes
-    dtype: string
-  - name: frame_nodes
-    dtype: string
-  - name: raw_content
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 16070511
-    num_examples: 33227
-  download_size: 6776171
-  dataset_size: 16070511
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
+    - name: id
+      dtype: string
+    - name: type
+      dtype: string
+    - name: name_nodes
+      dtype: string
+    - name: mean_nodes
+      dtype: string
+    - name: frame_nodes
+      dtype: string
+    - name: raw_content
+      dtype: string
+  num_rows: 33227
+license: agpl-3.0
+language:
+  - en
+  - vi
+tags:
+  - knowledge-graph
+  - ontology
+  - cognitive-architecture
+  - NMF
+size_categories:
+  - 10K<n<100K
 ---
+
+# Three-Reality Field Dataset
+
+33,227 grounded knowledge nodes organized into a spatial model of three reality layers.
+
+**Author**: Kevin T.N.
+
+## Structure
+
+Every node follows the NMF (Name-Meaning-Frame) schema:
+- **id** — unique hash
+- **type** — node category (existence, filter_gate, framework_layer, principle, discovery, topic, paper, text, code, ...)
+- **name_nodes** — JSON array of Name node IDs (what it is called)
+- **mean_nodes** — JSON array of Meaning node IDs (what it contains/connects to)
+- **frame_nodes** — JSON array of Frame node IDs (where it lives/what context)
+- **raw_content** — human-readable content
+
+## Three Realities
+
+The dataset models knowledge in three layers:
+- **R1 Subjective** — personal memory, unconscious patterns, layered center-to-surface
+- **R2 Intersubjective** — shared consensus, social knowledge (structural, not yet populated)
+- **R3 Absolute Objective** — what exists regardless of observers (the void outside)
+
+## Node Types
+
+| Type | Count | Description |
+|---|---|---|
+| doc_chunk | 21,560 | Content chunks from documents |
+| base_chunk_name | 9,002 | Chunk identifiers |
+| base_name | 1,098 | Entity names |
+| keyword | 498 | Extracted keywords |
+| paper | 391 | PDF/TeX papers |
+| text | 193 | Text files |
+| topic | 191 | Knowledge clusters |
+| document | 84 | DOCX/MD documents |
+| code | 80 | Python/JS code |
+| principle | 35 | General rules |
+| discovery | 20 | Insights |
+| design_spec | 18 | Architecture specs |
+| framework_layer | 6 | NMF, K/U, Collapse, IPOD, GAP, Superpose |
+| filter_gate | 5 | Binary, Gradient, Sequence, Relation, Recursion |
+| pyramid_vertex | 4 | Meta-methods: LMF, SDC(D/H)V, RT/OT, R1/R2/R3 |
+| existence | 1 | Root substrate (also = Gap) |
+
+## Grounding
+
+Every node is grounded to at least one of 5 primordial filter gates via its frame_nodes. Nodes closer to the center (existence) are more general. Nodes on the surface are more specific.
+
+## Usage
+
+
+
+## Related
+
+- [ThreeRealityField](https://github.com/jkdkr2439/ThreeRealityField) — 3D visualization + build tools
+- [ANC](https://github.com/jkdkr2439/ANC) — Anchor-Node Codec encoding pipeline
+- [QuadFusion](https://github.com/jkdkr2439/QuadFusion) — S/M/Q/R architecture experiment
+
+## License
+
+AGPL-3.0