{
 "id": "8Planetterraforming/Signal-Guided-Compression-Dataset-SGCD",
 "owner": "8Planetterraforming",
 "name": "Signal-Guided-Compression-Dataset-SGCD",
 "downloads": 7,
 "likes": 0,
 "last_modified": "2026-04-07T22:05:19.000Z",
 "stated_licence": "mit",
 "stated_from": "tag",
 "licence_name": null,
 "licence_link": null,
 "file_licence": null,
 "licence_files": [],
 "sources": [],
 "gated": false,
 "tasks": [],
 "citation": "@dataset{sgcd_v1,\n  title={Signal-Guided Compression Dataset},\n  author={8Planetterraforming},\n  year={2026},\n  publisher={Hugging Face}\n}\n---\n\n\n## Subsets\n\nThe SGCD dataset is composed of multiple structured subsets, each designed to enforce different aspects of consistency and signal-focused learning.\n\n### 1. Cube Multi-Object Consistency\n\nThis subset focuses on spatial consistency across multiple objects (e.g. cubes) and their faces.\n\nGoal:\n- enforce correct assignment between object surfaces and labels,\n- prevent duplication, omission, or misalignment,\n- train models to maintain structural integrity across multi-part inputs.\n\nKey properties:\n- multiple entities per sample,\n- positional relationships matter,\n- errors often occur as swapped or duplicated labels.\n\n---\n\n### 2. Alphabet Mapping (1–26)\n\nThis subset focuses on deterministic symbolic mapping.\n\nGoal:\n- enforce exact mapping between symbols and indices,\n- eliminate ambiguity in sequence relationships,\n- train models on clean, low-noise symbolic transformations.\n\nKey properties:\n- strict one-to-one mappings,\n- low entropy but high precision requirement,\n- sensitive to even small inconsistencies.\n\n---\n\n### 3. Geometric / Physical Consistency\n\nThis subset focuses on logical and physical correctness in structured scenarios.\n\nGoal:\n- ensure outputs follow basic geometric or physical rules,\n- detect contradictions in structured descriptions,\n- train models to respect underlying constraints.\n\nKey properties:\n- rule-based ",
 "signal": "permissive",
 "reason": "The declared licence (mit) is classified as permissive.",
 "revision": "bb0f608f6041ed42b980d923b16d9dab6f81fc3f",
 "stated_evidence": "https://huggingface.co/datasets/8Planetterraforming/Signal-Guided-Compression-Dataset-SGCD/blob/bb0f608f6041ed42b980d923b16d9dab6f81fc3f/README.md",
 "file_evidence": "https://huggingface.co/datasets/8Planetterraforming/Signal-Guided-Compression-Dataset-SGCD/tree/bb0f608f6041ed42b980d923b16d9dab6f81fc3f",
 "card_nc": false
}