quantum-physics-corpus-rebel-ready

Dataset Description

Quality-filtered corpus using classifier-based filtering (Phase 6/7 standard).

Filtering Criteria

Standard thresholds (Phase 6/7):

• domain_relevance >= 2 ("Clearly about quantum physics")
• technical_depth >= 2 ("Advanced undergraduate/graduate level")

Dataset Structure

Each chunk contains:

• text: The text content
• domain_relevance: Quality score (0-3)
• technical_depth: Quality score (0-3)
• structure_quality: Quality score (0-3)
• adu_types: Detected ADU types (list)
• adu_probs: ADU probabilities (dict)
• passes_quality_filter: True (all chunks passed)

Usage

from datasets import load_dataset

dataset = load_dataset("konsman/quantum-physics-corpus-rebel-ready")

for chunk in dataset['train']:
    print(f"Text: {chunk['text'][:100]}...")
    print(f"Quality: domain={chunk['domain_relevance']}, tech={chunk['technical_depth']}")
    print(f"ADU Types: {chunk['adu_types']}")

License

MIT License

Citation

Generated using Phase A quality filtering pipeline.