Breaking News Claim and Evidence Verification Dataset

#fact verification #claim-evidence matching #evidence-based classification #breaking news fact-checking #news credibility assessment #media information verification

5000
0.3GB
JSON
CATL
MOBIUSI INC

Updated:2026-11-03

AI Analysis & Value Prop

Designed for fact-checking breaking news webpages, this dataset pairs verifiable factual claims with source webpage information, relevant evidence spans, verdicts, and supporting rationales. Labels can indicate that evidence supports or contradicts a claim, or that the available evidence is insufficient. It supports evidence retrieval, claim-evidence matching, evidence-based judgment, and the development and evaluation of newsroom fact-checking workflows.

Dataset Insights

Sample Examples

{}

Technical Specifications

Field	Type	Description
file_name	string	File name
verdict	string	The conclusion reached about the claim after considering the webpage evidence.
claim_text	string	A verifiable factual claim that must be checked against webpage evidence.
evidence_spans	array	Evidence excerpts extracted from webpage content and their relationship to the claim.
source_webpage	object	The original webpage used as verification evidence, along with its source information.
verdict_rationale	string	An explanation of how the evidence supports the verdict and what information remains missing when evidence is insufficient.

Compliance Statement

Authorization Type	Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use	Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization	No PII, no real company names, simulated scenarios follow industry standards
Compliance System	Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

Frequently Asked Questions

What does the dataset contain?: Each record centers on a factual claim to be checked and includes source webpage information, relevant evidence spans, a verdict, and a rationale. Labels may indicate supporting, contradicting, or insufficient evidence; the exact label set depends on the released version.

Which model capabilities can it support?: The data can support training and evaluation for webpage evidence retrieval, claim-evidence matching, and evidence-based fact judgments. It can also assist newsroom verification workflows. Models should use source context and verdict rationales rather than making judgments from isolated text.

How can the data be obtained, and what format is provided?: Availability, delivery format, and dataset size should be confirmed on the release page or with the data provider. Before integration, review the field definitions, availability of webpage content, character encoding, and update schedule, and adapt ingestion and storage to the supplied evidence format.

What licensing and usage restrictions apply?: Refer to the license terms supplied by the data provider. In particular, verify the rights to reproduce, store, and redistribute source webpage text. Confirm the relevant permissions and privacy requirements before using the data for commercial training, public release, or downstream products.

Can't find the data you need?

Post a request and let data providers reach out to you.

Get this Dataset

Verified for Enterprise Use

Cite this Work

@dataset{Mobiusi2026,
  title={Breaking News Claim and Evidence Verification Dataset},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/b9cc91cf5058abe0d9095983b155078c?dataset_scene_id=17},
  urldate={2026-10-19},
  keywords={breaking news fact-checking dataset,news claim verification,claim-evidence matching,web evidence extraction,fact verification model,evidence retrieval training,media fact-checking data,claim truth assessment},
  version={1.0}
}

Using this in research? Please cite us.

placeholder

Scene

Domain

Modality

Task