Entropy (0–8): Higher ≈ more random. ≥7.0 often indicates compression/encryption/packing.

Compression (%): How much a block compresses. High % ≈ easier to compress (benign data). Very low or negative % with high entropy can indicate encryption/packing.

Uniformity: Max byte frequency / size. Very low ≈ uniform bytes (random-like); very high ≈ dominated by a few bytes (structured data).

Chi-square: Statistical deviation from uniform distribution. Large absolute changes (Δ) suggest structure change, boundaries, or packing stubs.

ML Confidence: Model certainty for the predicted class at that offset.