The Missing Parts: Augmenting Fact Verification with Half-Truth Detection

TL;DR

This paper introduces half-truth detection in fact verification, proposing a new benchmark and a modular framework that improves the identification of misleading omissions in political claims.

Contribution

It presents the novel task of half-truth detection, a new benchmark dataset, and a modular framework that enhances fact verification by modeling omitted information.

Findings

TRACER improves Half-True classification F1 by up to 16 points.

The benchmark includes 15k political claims with evidence alignment and inferred intent.

Modeling omissions significantly enhances fact verification accuracy.

Abstract

Fact verification systems typically assess whether a claim is supported by retrieved evidence, assuming that truthfulness depends solely on what is stated. However, many real-world claims are half-truths, factually correct yet misleading due to the omission of critical context. Existing models struggle with such cases, as they are not designed to reason about omitted information. We introduce the task of half-truth detection, and propose PolitiFact-Hidden, a new benchmark with 15k political claims annotated with sentence-level evidence alignment and inferred claim intent. To address this challenge, we present TRACER, a modular re-assessment framework that identifies omission-based misinformation by aligning evidence, inferring implied intent, and estimating the causal impact of hidden content. TRACER can be integrated into existing fact-checking pipelines and consistently improves performance across multiple strong baselines. Notably, it boosts Half-True classification F1 by up to 16 points, highlighting the importance of modeling omissions for trustworthy fact verification. The benchmark and code are available via https://github.com/tangyixuan/TRACER.