TQA-Bench: Evaluating LLMs for Multi-Table Question Answering with Scalable Context and Symbolic Extension

TL;DR

TQA-Bench is a comprehensive benchmark designed to evaluate large language models' ability to perform complex question answering over multi-table relational data, incorporating real-world datasets, scalable contexts, and symbolic reasoning extensions.

Contribution

We introduce TQA-Bench, a novel multi-table QA benchmark with scalable contexts and symbolic reasoning, addressing the limitations of existing single-table focused benchmarks.

Findings

LLMs show varying performance on multi-table QA tasks.

Symbolic extensions improve reasoning capabilities.

Larger models generally perform better on complex multi-table questions.

Abstract

The advent of large language models (LLMs) has unlocked great opportunities in complex data management tasks, particularly in question answering (QA) over complicated multi-table relational data. Despite significant progress, systematically evaluating LLMs on multi-table QA remains a critical challenge due to the inherent complexity of analyzing heterogeneous table structures and potential large scale of serialized relational data. Existing benchmarks primarily focus on single-table QA, failing to capture the intricacies of reasoning across multiple relational tables, as required in real-world domains such as finance, healthcare, and e-commerce. To address this gap, we present TQA-Bench, a new multi-table QA benchmark designed to evaluate the capabilities of LLMs in tackling complex QA tasks over relational data. Our benchmark incorporates diverse relational database instances sourced from real-world public datasets and introduces a flexible sampling mechanism to create tasks with varying multi-table context lengths, ranging from 8K to 64K tokens. To ensure robustness and reliability, we integrate symbolic extensions into the evaluation framework, enabling the assessment of LLM reasoning capabilities beyond simple data retrieval or probabilistic pattern matching. We systematically evaluate a range of LLMs, both open-source and closed-source, spanning model scales from 7 billion to 70 billion parameters. Our extensive experiments reveal critical insights into the performance of LLMs in multi-table QA, highlighting both challenges and opportunities for advancing their application in complex, data-driven environments. Our benchmark implementation and results are available at https://github.com/Relaxed-System-Lab/TQA-Bench.