Comparison

Matrix computation helpers used with visualization and cross-model analysis:

• compute_similarity_matrix — Pairwise similarity between models
• compute_grouped_similarity_matrices — Grouped similarity matrices (e.g. by layer or component)
• compute_gradiend_feature_cross_encoding_matrix — GRADIEND × feature-class cross-encoding matrix
• compute_gradiend_transition_cross_encoding_matrix — GRADIEND × transition matrix (pre-anchor)
• compute_anchor_aligned_encoding_matrix — Oriented feature × feature matrix
• compute_trainer_pair_encoding_matrix — Trainer × trainer matrix for binary positive-pair suites

See also the cross-model comparison guide and oriented cross-encoding matrix guide.