Datasets:

issdandavis
/

scbe-spine-overlay-proof-v1

configs:
  - config_name: default
    data_files:
      - split: train
        path: data/data.jsonl
language:
  - en
license: cc-by-4.0
pretty_name: SCBE Spine-Overlay Proof v1
size_categories:
  - n<1K
tags:
  - scbe
  - sacred-tongues
  - multi-domain
  - code
  - chemistry
  - robotics
  - motion
  - experimental

Status: experimental. Experiment-specific slice. Primary public dataset: scbe-aethermoore-training-data.

SCBE Spine-Overlay Proof v1

Tiny demonstration bundle (18 rows = 3 domains x 6 tongues) proving that the SCBE 12+ lane code-packet spine handles code, chemistry, and mechanical motion as overlays on a single tokenizer substrate, without forking the system.

Why this exists

Every row carries the same baseline lanes (binary, tokenizer, transport, labels, language_views, braille_lane, stisa, structural_parse, scip_symbol_index, semantic_token_bridge, route_ir, execution_lane, native_tokenization, atomic_states, ternary_semantics, semantic_expression).

Domain differentiation lives inside semantic_expression as a nested overlay:

Code rows carry no extra overlay — the 16 lanes already encode code structure.
Chemistry rows add chemistry_overlay (equation, reactants, products, reaction_class, stability, atoms_conserved, metaphor).
Motion rows add motion_assembly (platform_id, role, morphology_state, pilot_layers, comm_graph, embodiment_passport, invariants). Schema spec: docs/specs/MOTION_ASSEMBLY_SCHEMA.md.

Hard rules enforced

Both anchor_runtime and anchor_spirit are present on every row. Per-tongue map: KO=python, AV=typescript/javascript, RU=rust, CA=c/mathematica, UM=julia/haskell, DR=haskell/markdown.
ASCII-only field names and values. No Greek glyphs. CTBR vectors use omega_x/omega_y/omega_z, not Greek letters.
No branding or theming. Validated by tests/coding_spine/test_motion_assembly_schema.py.

Breakdown

domains: {'chemistry': 6, 'code': 6, 'motion': 6}
tongues: {'AV': 3, 'CA': 3, 'DR': 3, 'KO': 3, 'RU': 3, 'UM': 3}
row_count: 18

Files

data.jsonl — one JSON row per line, sorted keys, ASCII-encoded
manifest.json — counts, row ids, schema version
README.md — this file

Loading

from datasets import load_dataset
ds = load_dataset("issdandavis/scbe-spine-overlay-proof-v1")

Push

hf upload issdandavis/scbe-spine-overlay-proof-v1 \
  training-data/proofs/spine_overlay_proof_v1

Schema reference

docs/specs/MOTION_ASSEMBLY_SCHEMA.md — motion overlay spec
tests/coding_spine/test_motion_assembly_schema.py — schema test (28 cases)

License

CC-BY-4.0