NakedSignal OS · Documentation
Back to the consoleDocumentation

Census income

LIVE

The same instrument on census income prediction — the strongest split provenance in the suite, and the corpus where dimension five is measured against real demographic subgroups rather than a class proxy. No separate specification document has been published for this benchmark — the MedEval-1 spec is the document, and this page is that instrument run on another corpus. Everything below is read from the same payload.json the medical leaderboard renders from.

spec MedEval-1 v0.1code fingerprint de32bca72d56seed 20260727device mpsgenerated 2026-08-03T19:03+00:00
11
models scored
2
tasks
11
scored records
20,751
held-out cases
80.3
best index — GTE-base frozen + linear probe

The index is the weighted mean of the dimensions that were scored, on the weights published in the spec. Dimensions this corpus cannot support are excluded from it rather than counted as zero, which is why an index here is not comparable to a medical one model-for-model — it is a mean over a different set of questions, and the table below says which.

Every model, on all 9 dimensions

Ranked by index. An em-dash is a dimension the harness left null.
#ModelIndex1 Accuracy2 Cross-site3 Calibration4 Limited data5 Subgroup6 Corruption7 Uncertainty Specification sensitivity Cost, latency & footprint
1GTE-base frozen + linear probe109M frozen, mean-pooled 768-d, logistic head; contrastive sentence objective, general domain80.370.692.279.551.394.699.199.6
2TF-IDF + linear SVM (Platt-scaled)squared hinge, C=1, 5-fold Platt calibration78.476.398.372.525.096.899.8
3TF-IDF + MLP (256)SVD-256 then 1 hidden layer of 25675.373.097.835.740.097.299.8
4PubMedBERT-sentence frozen + linear probe109M frozen, mean-pooled 768-d, logistic head; biomedical pretraining PLUS a sentence objective75.271.695.363.522.493.198.699.8
5Gradient boosting200 stages, depth 3, lr 0.0572.555.991.680.649.385.793.1
6Logistic regressionstandardised, L2, C=170.653.397.179.129.490.691.1
7BiomedBERT frozen + linear probe109M frozen, mean-pooled 768-d, logistic head69.966.192.273.20.086.498.798.9
8TF-IDF + logistic regression1-2 grams, 200k features, sublinear tf69.659.493.884.80.093.599.7
9Random forest400 trees69.452.885.090.039.283.889.6
10MLP (128, 64)2 hidden layers, adam, early stopping when n>=20069.150.892.387.724.989.790.7
11TF-IDF + complement naive Bayes1-2 grams, alpha=0.367.752.588.787.00.098.199.3
Empty columns, and why they are empty
  • 2. Cross-site robustness No re-acquisition of the same subjects exists in this corpus, so there is nothing to measure. Widening the definition to fill the cell would be a spec change, not a demo.
  • . Not computed on any record in this benchmark, and not approximated.

The measurements the scores came from

Raw metrics, not the 0–100 rescaling.
ModelBalanced accuracy (95% CI)AUCECERisk–coverage AUCMean label-budget retention
GTE-base frozen + linear probe0.853[0.8240.882]0.9490.0160.9950.795
TF-IDF + linear SVM (Platt-scaled)0.882[0.8550.909]0.9860.0030.9990.725
TF-IDF + MLP (256)0.865[0.8380.893]0.9830.0040.9990.357
PubMedBERT-sentence frozen + linear probe0.858[0.8250.884]0.9310.0090.9930.635
Gradient boosting0.779[0.7730.789]0.9190.0170.9650.806
Logistic regression0.766[0.7590.775]0.9010.0060.9550.791
BiomedBERT frozen + linear probe0.831[0.7980.859]0.9360.0160.9940.732
TF-IDF + logistic regression0.797[0.7620.829]0.9840.0120.9980.848
Random forest0.764[0.7570.772]0.8890.0300.9480.900
MLP (128, 64)0.754[0.7470.763]0.8970.0150.9530.877
TF-IDF + complement naive Bayes0.763[0.7300.794]0.9490.0230.9960.870

Balanced accuracy carries a bootstrap 95% interval; read it before reading the rank. ECE is expected calibration error, so lower is better and the 0–100 dimension score inverts it. Every value here is on the record; hover any figure for the unrounded number.

Who the best model fails

LIVE

Dimension five, on GTE-base frozen + linear probe — the top-ranked model above. Basis, in the evaluator’s own words: worst demographic subgroup (real metadata). This is the strong reading — the corpus ships the attributes, so the harness did not have to fall back to the worst-class proxy every medical imaging task uses.

countrygap 0.230 · worst group 0.756
AUworst0.756
unknown0.800
US0.852
CA0.875
GB0.912
IN0.986
required_educationgap 0.206 · worst group 0.794
High School or equivalentworst0.794
not specified0.852
Unspecified0.859
Bachelor's Degree0.876
Certification1.000
Master's Degree1.000
Professional1.000
Some College Coursework Completed1.000

Per-group recall, and the gap between the best and worst group. These numbers measure who the model fails on this corpus. They are not an audit of any deployed system, and the subgroup attributes are model inputs here, as they are in the standard formulation of the task.

The corpus

Census income (>$50K)

adult_income · Tabular / socioeconomic
Source
UCI Adult / Census IncomeKohavi 1996 / UCI ML Repository, Adult
Split
official released split (UCI adult.test scored whole); validation carved from adult.data only (stratified 15%, seed 20260727)
Sizes
12,000 train · 16,281 held out · 2 classesTraining capped at the standard budget from 27,676 available rows, so every model in the suite sees the same amount of data.
Classes
<=50K · >50K
Registry note
Official released train/test split -- the test file is used whole and validation is carved from the official train file only. Carries sex, race and age, so dimension 5 is measured against real demographic subgroups rather than the worst-class fallback. Dimension 2 (cross-site robustness) is undefined: there is no re-acquisition of the same subjects to measure. Known critique: Ding et al., "Retiring Adult: New Datasets for Fair Machine Learning" (NeurIPS 2021) argues this 1994 census extract is outdated as a fairness benchmark and proposes ACS/folktables as the replacement; that task is on the roadmap and this one is published with the criticism attached, not without it.

Fraudulent job posting detection

jobpost_fraud · Recruitment text
Source
EMSCAD (Employment Scam Aegean Dataset)Vidros, Kolias, Kambourakis & Akoglu, Future Internet 9(1):6, 2017
Split
no official split published; 60/15/25 train/val/test minted at the standard seed (20260727), stratified on the 4.8% positive class
Sizes
10,728 train · 4,470 held out · 2 classes
Classes
genuine · fraudulent
Registry note
4.8% of 17,880 postings are fraudulent, so a 25% held-out split carries roughly 215 positives and the intervals are wide by construction. LICENCE: the HuggingFace mirror tags this CC0, but the underlying EMSCAD release from the University of the Aegean is CC BY-NC-SA 4.0. It is treated as non-commercial here and cited to the original. Subgroups are recorded fields (country, required education), not inferred attributes.

Labour market

The same dimensions, pointed at census income prediction.

The census income task is scored by the identical harness that scores the medical suite — same evaluators, same seed, same chance-correction, same record format. Nothing in the scoring code knows what a patient is, which is why adding a domain is a registry row and an adapter branch rather than a second product.

It is also the strongest split provenance in the whole suite. The corpus ships an official train/test split, so the held-out file is scored whole and validation is carved from the training file alone — no minted split, no seed to trust. Every task card states which of the two it is.

And because the corpus carries sex, race and age, dimension five is measured against real demographic subgroups here rather than the worst-class proxy the imaging tasks fall back to.

What this does not cover
  • The corpus is a 1994 census extract and has a published critique attached: Ding et al., "Retiring Adult: New Datasets for Fair Machine Learning" (NeurIPS 2021) argues it is outdated as a fairness benchmark and proposes ACS/folktables as the replacement. That task is on the roadmap. This one ships with the criticism cited on its card, which is the only version of shipping it that a standard can defend.
  • Dimension two is undefined here — there is no re-acquisition of the same subjects to measure — and renders as an em-dash rather than a zero. Dimension six is measured: field dropout, coding drift, unit change and entry error, at three severities. Coding drift is reported unscored on this corpus, because its categoricals arrive pre-encoded as ordinals and the transform has no one-hot block to act on.
  • The subgroup attributes are also model inputs, as they are in the standard formulation of this task. The dimension-five numbers measure who the model fails; they are not an audit of a deployed lending or hiring system.

Re-derive every figure on this page with python src/build_site.py --emit-json in plan_d/ — the same run writes the medical leaderboard. All benchmarks reads the same 9 dimensions per domain, and the trust graph traces any figure here back to the run that produced it.