Latest Benchmark

Rolling nightly benchmark compiled from the live operational runs. Snapshot label: latest-2026-05-19T2133Z.

Trained params is the average non-zero parameter count in the fitted model across folds. Avg active params / classification is the mean number of active parameters or scoring decisions touched while classifying one product.

11,538Filtered products

3,818Filtered tags

528Filtered taxonomies

79Prime base

View UMLLR tag-order ablation View operational UMLLR page

Benchmark dashboard chart for latest benchmark view — Small-multiples dashboard generated directly from the comparison table rows below.

Model comparison chart for latest benchmark view — Log-log scatter of trained parameters versus mean p-adic loss, overlaid with the current parsimoniousness baseline.

Active parameter comparison chart for latest benchmark view — Scatter plot of avg active params versus mean p-adic loss, excluding PCLR and PCNN. Fitted equation: mean p-adic loss = -0.0912 log10(active params) + 0.4030. Regression fitted on log10(active params): R²=0.655, p=0.028.

Active parameter versus log-loss chart for latest benchmark view — Scatter plot of avg active params versus log10(mean p-adic loss), excluding PCLR and PCNN. Fitted equation: log10(mean p-adic loss) = -0.1546 log10(active params) - 0.4216. Regression fitted on log10(active params): R²=0.803, p=0.006.

Model	Trained params	Mean p-adic loss	Exact acc.	Prefix-2 acc.	Avg active params / classification
Dummy Baseline	1.0	0.601667	4.22%	19.75%	1.00
Importance-Optimised p-adic Linear Regression	1049.0	0.288520	48.50%	62.81%	1.10
Parameter-constrained Logistic Regression	17344.8	0.783546	5.03%	15.97%	753.98
Unconstrained Logistic Regression with L1	6581.0	0.113444	63.79%	80.02%	542.71
Parameter-constrained Neural Network	15607.8	0.536307	12.13%	29.36%	14755.53
Decision Tree	55009.6	0.148222	56.75%	75.90%	390.46
Level-wise Logistic Regression	212198.6	0.136510	52.64%	76.71%	213.87
Unconstrained Neural Network with L1	52660.8	0.129984	61.31%	78.65%	6905.46
Zubarev (greedy init.)	1195.0	0.337137	42.95%	58.12%	1.41