ContextClaim: A Context-Driven Paradigm for Verifiable Claim Detection

Yufeng Li; Rrubaa Panchendrarajan; Arkaitz Zubiaga

cs.CLMar 31, 2026

ContextClaim: A Context-Driven Paradigm for Verifiable Claim Detection

Authors: Yufeng Li, Rrubaa Panchendrarajan, Arkaitz Zubiaga

Organizations: School of Electronic Engineering and Computer Science, Queen Mary University of London, London E1 4NS, U.K.

Abstract

Automated fact-checking pipelines typically begin with a filtering stage that decides which claims are worth verifying, given that the later evidence retrieval and verification components are expensive to apply at scale. A central task in this stage is verifiable claim detection, which asks whether a statement is in principle checkable against external evidence. Prior work on this task, as well as on the closely related notion of check-worthiness, conditions its decisions only on the claim sentence itself. We argue that this is restrictive, because deciding whether a statement is checkable often depends on identifying the entities and events it mentions, and on whether external information about them is actually available in the first place. Motivated by how downstream verification systems rely on retrieved evidence, we move retrieval upstream into the detection stage and introduce ContextClaim. Given an input claim, the approach identifies entity mentions, queries Wikipedia as a structured background source, and uses large language models to compress the retrieved material into short contextual summaries that are then passed to a classifier. Experiments are conducted on two domains and genres, namely the CheckThat! 2022 Twitter collection and the PoliClaim corpus of political debates, and cover both encoder and decoder only models under fine-tuning, zero-shot, and few-shot settings. The added context yields gains on verifiable claim detection in several configurations, although the size of the improvement varies with the dataset, the backbone model, and the training setup. We further find that the same retrieved summaries are useful beyond detection. Feeding them into a downstream verification model on FEVER improves verification F1. Component level analyses, human annotation, and error inspection further clarify the conditions under which retrieved context helps, and where it does not.

Figures & tables

Fig. 1: Example of ContextClaim system processing a vaccine claim about Lindsey Graham.

Split	CT22			PoliClaim †
Split	Total	Veri	Non	Total	Veri	Non
Train	3,324	64%	36%	1,757	59%	41%
Dev	307	64%	36%	196	59%	41%
Test	251	59%	41%	816	64%	36%

TABLE I: Dataset Statistics. Veri = Verifiable, Non = Non-verifiable.

		CT22		PoliClaim
Model	System	Acc	F1	Acc	F1
BERT	Baseline	68.29 ${}_{\pm\text{1.37}}$	76.45 ${}_{\pm\text{0.72}}$	82.77 ${}_{\pm\text{0.41}}$	87.07 ${}_{\pm\text{0.32}}$
	CC-G4o	70.20 ${}_{\pm\text{1.77}}$	77.18 ${}_{\pm\text{1.06}}$	84.02 ${}_{\pm\text{0.97}}$	87.94 ${}_{\pm\text{0.59}}$
	CC-M	69.48 ${}_{\pm\text{0.83}}$	76.89 ${}_{\pm\text{0.28}}$	84.09 ${}_{\pm\text{1.48}}$	88.06 ${}_{\pm\text{0.81}}$
RoBERTa	Baseline	70.84 ${}_{\pm\text{1.94}}$	78.53 ${}_{\pm\text{0.50}}$	85.78 ${}_{\pm\text{0.92}}$	89.42 ${}_{\pm\text{0.67}}$
	CC-G4o	71.95 ${}_{\pm\text{3.40}}$	78.64 ${}_{\pm\text{1.85}}$	86.45 ${}_{\pm\text{0.53}}$	89.71 ${}_{\pm\text{0.20}}$

TABLE II: Fine-tuning results on CT22 and PoliClaim (mean

\pm

std). Bold: column-wise best. Underline: overall best per metric.

		CT22				PoliClaim
		Zero-shot		Few-shot		Zero-shot		Few-shot
Model	System	Acc	F1	Acc	F1	Acc	F1	Acc	F1
Llama3	Baseline	60.16	74.87	71.31	79.78	69.24	79.71	72.55	80.42
	CC-G4o	62.15	75.83	68.53	78.59	70.96	80.23	74.26	81.87
	CC-M	62.55	76.02	67.33	78.07	71.81	80.87	74.51	80.74
Mistral	Baseline	67.73	78.28	63.75	76.49	79.66	84.28	71.08	81.21

TABLE III: Zero-shot and few-shot results on CT22 and PoliClaim. Bold: column-wise best. Underline: overall best per metric.

System	Dataset	Acc (%)
Majority Baseline	CT22	59.36
CheckThat!’22 Best [ 24 ]	CT22	76.10
ContextClaim CC-G4o (Ours)	CT22	79.28
Majority Baseline	PoliClaim	63.85
AFaCTA [ 12 ]	PoliClaim	86.27
ContextClaim CC-G4o (Ours)	PoliClaim	86.45

TABLE IV: Comparison with published systems on CT22 and PoliClaim.

Fig. 2: Performance Improvement on two Datasets with three Settings.

Fig. 3: Entity Coverage and Quantity Comparison of PoliClaim and CT22.

Fig. 4: Precision-Recall Trade-off on Two Datasets.

Fig. 5: Model Architecture Effects on Contextual Information Benefits.

Fig. 6: Learning Setting Comparison.

CT22
Setting	Model	BL $\rightarrow$ Raw		Raw $\rightarrow$ G4o		Raw $\rightarrow$ M
Setting	Model	Acc	F1	Acc	F1	Acc	F1
FT	Llama3	-0.80	+1.73	+0.80	+0.40	+1.06	+0.79
	Mistral	+11.02	+3.61	+1.06	+0.21	-2.92	-3.20
ZS	Llama3	+3.19	+1.42	-1.20	-0.46	-0.80	-0.27
	Mistral	-0.80	-0.65	-1.59	-0.59	-1.59	-0.59

TABLE V: Performance Gain (

\Delta

) from ContextClaim Components.

Dimension	GPT-4o	Mistral	Diff	p-value
Relevance	2.93 $\pm$ 0.03	2.88 $\pm$ 0.04	+0.06	0.21
Signal Clarity	1.80 $\pm$ 0.08	1.74 $\pm$ 0.07	+0.06	0.41
Usefulness	1.62 $\pm$ 0.07	1.58 $\pm$ 0.07	+0.03	0.59
Overall	2.12 $\pm$ 0.90	2.07 $\pm$ 0.89	+0.05	–

TABLE VI: Human evaluation results for context summary quality. Ratings on 1–3 scale (1=Poor, 2=Acceptable, 3=Good). No significant differences between systems (paired t-test, all

p>0.05

Model	FP	FN	FP/FN	Errors
Llama3 Baseline	71 (28.3%)	23 (9.2%)	3.09	94 (37.5%)
Llama3 + CC-G4o	86 (34.3%)	7 (2.8%)	12.29	93 (37.1%)
RoBERTa Baseline	63 (25.1%)	13 (5.2%)	4.85	76 (30.3%)
RoBERTa + CC-M	48 (19.1%)	18 (7.2%)	2.67	66 (26.3%)

TABLE VII: Error Distribution by Model on CT22 (n=251).

Metric	Claim Only	Claim + CC-G4o	$\boldsymbol{\Delta}$
Accuracy	88.70	89.34	+0.64
Precision	91.00	87.97	$-$ 3.03
Recall	85.89	91.14	+5.25
F1	88.37	89.53	+1.16
Macro-F1	88.69	89.34	+0.65

TABLE VIII: Downstream verification on FEVER (SUPPORTS vs. REFUTES,

n{=}6{,}666

). GPT-4o classifier is held fixed; only the input varies. F1: positive=SUPPORTS.

Automated Fact-Checking

Automated fact-checking detects checkable claims, retrieves evidence, and assesses whether that evidence supports or refutes them.

See all 139 papers

CommunityFact: A Dynamic, Multilingual, Multi-domain Benchmark for Misinformation Detection in the Wild
May 28, 2026Sahajpreet Singh, Insyirah Mujtahid, Min-Yen Kan +1Multilingual Language Model Evaluation LLM Evaluation
National University of Singapore
Hybrid Fact-Checking that Integrates Knowledge Graphs, Large Language Models, and Search-Based Retrieval Agents Improves Interpretable Claim Verification
Nov 5, 2025Shaghayegh Kolli, Richard Rosenbaum, Timo Cavelius +3LLM Grounding Automated Fact-Checking
Technical University Munich
ThinknCheck: Grounded Claim Verification with Compact, Reasoning-Driven, and Interpretable Models
Apr 2, 2026Delip Rao, Feijiang Han, Chris Callison-BurchClaim Verification LLM Answer Verification
University of Pennsylvania, Philadelphia, PA, USA

May 28, 2026J/K move · Enter open · S save

CommunityFact: A Dynamic, Multilingual, Multi-domain Benchmark for Misinformation Detection in the Wild

Sahajpreet Singh, Insyirah Mujtahid, Min-Yen Kan, Kokil Jaidka

National University of Singapore

Multilingual Language Model Evaluation LLM Evaluation Fake News Detection Automated Fact-Checking LLM Reliability

Open paper AskPDF arXiv

Misinformation verification increasingly occurs in public, fast-moving, and multilingual online settings, where static benchmarks provide an incomplete measure of model reliability. We introduce CommunityFact, a refreshable benchmark for misinformation detection in the wild, with three major goals: coverage, granularity, and redistributability. This release contains 15,992 standalone claims across five languages and two domains. We evaluate ten LLMs under varying inference-time capabilities, including thinking and web-search. Our results show that closed-input verification remains challenging, web access yields the largest gains, and web-enabled LLMs' source-selection policies are systematically misaligned with the sources human Community Notes raters converge on -- a gap that closes through model-specific mechanisms of retrieval expansion or pruning. We further find substantial variation across language-domain slices and across the evidence ecosystems used by web-enabled systems. Beyond evaluation, CommunityFact positions Community Notes as a training signal for claim-conditioned source suggesters that could improve factual verification on novel claims.

Abstract

Automated Fact-Checking

Explore similar work

CommunityFact: A Dynamic, Multilingual, Multi-domain Benchmark for Misinformation Detection in the Wild

Hybrid Fact-Checking that Integrates Knowledge Graphs, Large Language Models, and Search-Based Retrieval Agents Improves Interpretable Claim Verification

ThinknCheck: Grounded Claim Verification with Compact, Reasoning-Driven, and Interpretable Models