Are Large Language Models Table-based Fact-Checkers?

TL;DR

This paper investigates the potential of large language models for table-based fact verification, demonstrating their effectiveness with prompt engineering and instruction tuning, and providing insights for future table reasoning research.

Contribution

It is the first to systematically explore LLMs' zero-shot and few-shot capabilities in TFV and how instruction tuning enhances their performance.

Findings

LLMs achieve acceptable zero-shot and few-shot TFV results with prompt engineering.

Instruction tuning significantly improves LLMs' TFV performance.

Format of prompts and number of examples influence TFV accuracy.

Abstract

Table-based Fact Verification (TFV) aims to extract the entailment relation between statements and structured tables. Existing TFV methods based on small-scaled models suffer from insufficient labeled data and weak zero-shot ability. Recently, the appearance of Large Language Models (LLMs) has gained lots of attraction in research fields. They have shown powerful zero-shot and in-context learning abilities on several NLP tasks, but their potential on TFV is still unknown. In this work, we implement a preliminary study about whether LLMs are table-based fact-checkers. In detail, we design diverse prompts to explore how the in-context learning can help LLMs in TFV, i.e., zero-shot and few-shot TFV capability. Besides, we carefully design and construct TFV instructions to study the performance gain brought by the instruction tuning of LLMs. Experimental results demonstrate that LLMs can achieve acceptable results on zero-shot and few-shot TFV with prompt engineering, while instruction-tuning can stimulate the TFV capability significantly. We also make some valuable findings about the format of zero-shot prompts and the number of in-context examples. Finally, we analyze some possible directions to promote the accuracy of TFV via LLMs, which is beneficial to further research of table reasoning.