Contamination Map Methodology

Tabular / Model

LimiX

Organisation
Stable AI Technology Co., Ltd. (LimiX team)
Family
LimiX · version v1 (LimiX-16M, LimiX-2M)
Released
2025-08-29 (Model card news entry '[2025.08.29] LimiX V1.0 Released.' https://huggingface.co/stable-ai/LimiX-16M (LimiX-2M released 2025-11-10))
Checkpoints
stable-ai/LimiX-16M, stable-ai/LimiX-2M, stable-ai/LimiX-1_16M, stable-ai/LimiX-1_2M
Weights
Open · Paper: Apache 2.0; HF weights: limix-16m-license-v1.0 (custom)
Links
PaperModel cardCode

Training data disclosure: Full list

Level determined for 100% of benchmark cells.

“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Other declared training data

Benchmark cells

TabArena

Not seen airfoil_self_noise · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen Amazon_employee_access · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen anneal · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen Another-Dataset-on-used-Fiat-500 · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen APSFailure · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen bank-marketing · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen Bank_Customer_Churn · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen Bioresponse · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen blood-transfusion-service-center · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen churn · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen coil2000_insurance_policies · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen concrete_compressive_strength · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen credit-g · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen credit_card_clients_default · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen customer_satisfaction_in_airline · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen diabetes · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen Diabetes130US · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen diamonds · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen E-CommereShippingData · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen Fitness_Club · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen Food_Delivery_Time · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen GiveMeSomeCredit · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen hazelnut-spread-contaminant-detection · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen healthcare_insurance_expenses · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen heloc · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen hiva_agnostic · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen houses · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen HR_Analytics_Job_Change_of_Data_Scientists · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen in_vehicle_coupon_recommendation · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen Is-this-a-good-customer · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen kddcup09_appetency · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen Marketing_Campaign · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen maternal_health_risk · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen miami_housing · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen MIC · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen NATICUSdroid · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen online_shoppers_intention · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen physiochemical_protein · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen polish_companies_bankruptcy · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen qsar-biodeg · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen QSAR-TID-11 · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen QSAR_fish_toxicity · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen SDSS17 · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen seismic-bumps · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen splice · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen students_dropout_and_academic_success · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen superconductivity · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen taiwanese_bankruptcy_prediction · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen website_phishing · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen wine_quality · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Not seen jm1 · absent from the complete list of training data

Evidence (1)
“The pretraining data for LimiX models is generated with hierarchical SCMs via graph-aware and solvability-aware sampling.”
arxiv.org · paper · Conclusion · retrieved 2026-10-15

Notes

LimiX-1_16M / LimiX-1_2M (Sept 2026) appear to be renamed re-uploads of the v1 models (card differs only in license name); LimiX-2M has a separate paper (arXiv 2606.04485) not read here.

Sources