mlsanityData to 7 Oct 2026

Marginal Response Surface Elicitation for Zero-Label Tabular Learning

arXiv 2609.39639 ↗First posted 30 Sep 2026cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

Tabular learning uses structured data to predict target outcomes. Traditionally, this process has relied on labeled data. However, large language models (LLMs) can be used to elicit domain priors based on the task description and feature semantics, thereby enabling predictions without labeled data. We propose Marginal Response Surface Elicitation (MARS), a method that transforms feature-level LLM priors into a reusable, zero-shot tabular classifier. To construct this classifier, MARS selects representative values for each feature from unlabeled data and prompts the LLM to provide corresponding class support scores and feature weights. It then aggregates multiple responses using the median to construct feature response functions, and makes predictions through their weighted sum without further LLM queries. Across eight tabular benchmark tasks, MARS achieves the highest average AUC and AP, outperforming direct prompting by 1.97 and 6.21 percentage points respectively, while substantially reducing end-to-end costs. Evaluations with LLMs of different sizes further demonstrate its predictive advantage over direct prompting.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 20 of its 40 references that are in the map. First posted on arXiv 30 Sep 2026.

Builds on

20 of 40 references.

References · 20

  1. 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,120
  2. 2021Deep Neural Networks and Tabular Data: A Surveycited by 489
  3. 2022TabLLM: Few-shot Classification of Tabular Data with Large Language Modelscited by 273
  4. 2025TabArena: A Living Benchmark for Machine Learning on Tabular Datacited by 152
  5. 2021SCARF: Self-Supervised Contrastive Learning using Random Feature Corruptioncited by 151
  6. 2026TabICLv2: A better, faster, scalable, and open tabular foundation modelcited by 112
  7. 2024TabM: Advancing Tabular Deep Learning with Parameter-Efficient Ensemblingcited by 105
  8. 2024Large Language Models Can Automatically Engineer Features for Few-Shot Tabular Learningcited by 79
  9. 2024Large Scale Transfer Learning for Tabular Data via Language Modelingcited by 73
  10. 2023Curated LLM: Synergy of LLMs and Data Curation for tabular augmentation in low-data regimescited by 52
  11. 2023STUNT: Few-shot Tabular Learning with Self-generated Tasks from Unlabeled Tablescited by 48
  12. 2024Elephants Never Forget: Memorization and Learning of Tabular Data in Large Language Modelscited by 34
8 more
  1. 2023TABLET: Learning From Instructions For Tabular Datacited by 28
  2. 2024AutoElicit: Using Large Language Models for Expert Prior Elicitation in Predictive Modellingcited by 15
  3. 2020Trajectories, bifurcations and pseudotime in large clinical datasets: applications to myocardial infarction and diabetes datacited by 13
  4. 2025LLM Meeting Decision Trees on Tabular Datacited by 7
  5. 2024"Oh LLM, I'm Asking Thee, Please Give Me a Decision Tree": Zero-Shot Decision Tree Induction and Embedding with Large Language Modelscited by 6
  6. 2024Small Models are LLM Knowledge Triggers on Medical Tabular Predictioncited by 4
  7. 2025Latte: Transfering LLMs` Latent-level Knowledge for Few-shot Tabular Learningcited by 3
  8. 2025LLM Empowered Prototype Learning for Zero and Few-Shot Tasks on Tabular Datacited by 2

Built on by

Not cited yet.