TABVERSE benchmark shows representation format substantially affects LLM and VLM performance on table QA, structural understanding, and reconstruction tasks.
arXiv preprint arXiv:2411.02059 , year =
12 Pith papers cite this work, alongside 4 external citations. Polarity classification is still indexing.
12
Pith papers citing it
4
external citations · Pith
abstract
The emergence of models like GPTs, Claude, LLaMA, and Qwen has reshaped AI applications, presenting vast new opportunities across industries. Yet, the integration of tabular data remains notably underdeveloped, despite its foundational role in numerous real-world domains. This gap is critical for three main reasons. First, database or data warehouse data integration is essential for advanced applications; second, the vast and largely untapped resource of tabular data offers immense potential for analysis; and third, the business intelligence domain specifically demands adaptable, precise solutions that many current LLMs may struggle to provide. In response, we introduce TableGPT2, a model rigorously pre-trained and fine-tuned with over 593.8K tables and 2.36M high-quality query-table-output tuples, a scale of table-related data unprecedented in prior research. This extensive training enables TableGPT2 to excel in table-centric tasks while maintaining strong general language and coding abilities. One of TableGPT2's key innovations is its novel table encoder, specifically designed to capture schema-level and cell-level information. This encoder strengthens the model's ability to handle ambiguous queries, missing column names, and irregular tables commonly encountered in real-world applications. Similar to visual language models, this pioneering approach integrates with the decoder to form a robust large multimodal model. We believe the results are compelling: over 23 benchmarking metrics, TableGPT2 achieves an average performance improvement of 35.20% in the 7B model and 49.32% in the 72B model over prior benchmark-neutral LLMs, with robust general-purpose capabilities intact.
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
TRL-Bench is a new multi-granular benchmark that releases 50 OpenML tables, linkage tasks, and a 47k-table data lake to show that tabular encoder performance is capability-specific rather than captured by one leaderboard.
MACE deploys three specialized LLM agents (Planner, Executor, Verifier) with zero-shot CoT to verify claims from tables, matching SOTA performance on two datasets and near-SOTA on two others using models 2-8x smaller than prior bests.
DataGovBench is a new benchmark using 178 large multi-tabular government datasets showing state-of-the-art LLMs and agents achieve below 40% QA accuracy and below 50% insight scores, far from real-world data analysis demands.
Empirical 2x2 factorial study on 6 statistical datasets shows format and schema constraints in LLM-based KG construction from CSV tables produce super-additive fidelity loss up to +1.180, with mismatched pairs falling below baseline, plus release of CSVFidelity-Bench.
Semantically invariant row and column permutations in tables can cause LLMs to output incorrect answers, and a gradient-based attack called ATP efficiently finds such permutations that degrade performance across many models.
SpreadsheetAgent uses incremental multi-format reading, structural sketching, and verification to raise spreadsheet benchmark accuracy from 35.27% to 38.16%.
RedParrot accelerates NL-to-DSL conversion by 3.6x with 8.26% accuracy gain on enterprise data and 34.8% on benchmarks via semantic caching of query skeletons and contrastive learning.
MachineLearningLM uses continued pretraining on SCM-synthesized ML tasks with random-forest distillation to give LLMs robust many-shot in-context learning on tabular classification, reaching random-forest accuracy levels while preserving general chat performance.
MoCA-Agent decomposes questions into typed atomic claims, clears them via trader-agent markets into confidence-weighted decisions, synthesizes and verifies executable Python code, and reports strong benchmark scores including 78.3% on FinQA.
Constructs synthetic contrastive reasoning-trace dataset for MMQA via heterogeneous LLMs and reports 9.7-16.3% absolute gains from CPO fine-tuning over standard SFT across three open LLMs.
citing papers explorer
-
TABVERSE: Benchmarking Cross-Format Table Understanding in LLMs and VLMs
TABVERSE benchmark shows representation format substantially affects LLM and VLM performance on table QA, structural understanding, and reconstruction tasks.
-
TRL-Bench: Standardizing Cross-Paradigm Representation-Level Evaluation of Tabular Encoders
TRL-Bench is a new multi-granular benchmark that releases 50 OpenML tables, linkage tasks, and a 47k-table data lake to show that tabular encoder performance is capability-specific rather than captured by one leaderboard.
-
A Multi-Agent Approach for Claim Verification from Tabular Data Documents
MACE deploys three specialized LLM agents (Planner, Executor, Verifier) with zero-shot CoT to verify claims from tables, matching SOTA performance on two datasets and near-SOTA on two others using models 2-8x smaller than prior bests.
-
Data Analysis in the Wild: Benchmarking Large Language Models Against Real-World Data Complexities
DataGovBench is a new benchmark using 178 large multi-tabular government datasets showing state-of-the-art LLMs and agents achieve below 40% QA accuracy and below 50% insight scores, far from real-world data analysis demands.
-
Format-Constraint Coupling in Knowledge Graph Construction from Statistical Tables
Empirical 2x2 factorial study on 6 statistical datasets shows format and schema constraints in LLM-based KG construction from CSV tables produce super-additive fidelity loss up to +1.180, with mismatched pairs falling below baseline, plus release of CSVFidelity-Bench.
-
The Power of Order: Fooling LLMs with Adversarial Table Permutations
Semantically invariant row and column permutations in tables can cause LLMs to output incorrect answers, and a gradient-based attack called ATP efficiently finds such permutations that degrade performance across many models.
-
Towards Robust Real-World Spreadsheet Understanding with Multi-Agent Multi-Format Reasoning
SpreadsheetAgent uses incremental multi-format reading, structural sketching, and verification to raise spreadsheet benchmark accuracy from 35.27% to 38.16%.
-
RedParrot: Accelerating NL-to-DSL for Business Analytics via Query Semantic Caching
RedParrot accelerates NL-to-DSL conversion by 3.6x with 8.26% accuracy gain on enterprise data and 34.8% on benchmarks via semantic caching of query skeletons and contrastive learning.
-
MachineLearningLM: Scaling Many-shot In-context Learning via Continued Pretraining
MachineLearningLM uses continued pretraining on SCM-synthesized ML tasks with random-forest distillation to give LLMs robust many-shot in-context learning on tabular classification, reaching random-forest accuracy levels while preserving general chat performance.
-
MoCA-Agent: A Market-of-Claims Code Agent for Financial and Numerical Reasoning
MoCA-Agent decomposes questions into typed atomic claims, clears them via trader-agent markets into confidence-weighted decisions, synthesizes and verifies executable Python code, and reports strong benchmark scores including 78.3% on FinQA.
-
Synthetic Contrastive Reasoning for Multi-Table Q&A
Constructs synthetic contrastive reasoning-trace dataset for MMQA via heterogeneous LLMs and reports 9.7-16.3% absolute gains from CPO fine-tuning over standard SFT across three open LLMs.
- Efficient Table QA via TableGrid Navigation and Progressive Inference Prompting