Breaking News Event Timeline Reconstruction Question-Answering Dataset

#Multi-hop Question Answering #Timeline Reconstruction #Evidence Retrieval #Uncertainty Identification #Breaking News Analysis #Event Progress Tracking #News Fact Checking #Media Content Understanding
Records
4710
Size
0.2GB
Format
JSON
Provider
MOBIUSI INCMOBIUSI INC
Updated
v1.02026-11-04
AI Analysis & Value Prop

This dataset brings together multiple news reports or consecutive updates about the same breaking event. Its question-answer examples require linking information across texts, including times, actions, and changes in event status, and provide answers, reconstructed timelines, and supporting evidence. Samples also capture ambiguities, conflicting accounts, or unresolved details, making the collection useful for evidence-grounded multi-hop reasoning, reasoning-focused supervised fine-tuning, and evaluation of news event understanding.

Overview

Dataset Insights

Preview

Sample Examples

sample.json
{}

Schema

Technical Specifications

file_namestringFile name
answerstringA concise, source-grounded response to the question.
evidencearrayEvidence drawn from the sources to support the answer and timeline reconstruction.
questionstringA question that requires linking information across the source reports and updates.
timelinearrayEvents reconstructed from the source reports and arranged in chronological order.
uncertaintiesarrayAmbiguities, conflicting reports, or unresolved details that limit the reconstruction.
source_reportsstringOriginal articles or consecutive updates about the breaking event, including available source labels and timestamps.

Trust

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

FAQ

Frequently Asked Questions

What does the dataset contain?
Each example focuses on a breaking event and includes source reports or consecutive updates, a question that links information across texts, a reference answer, a chronologically ordered timeline, and supporting evidence. Some examples also record conflicts, ambiguities, or unresolved details.
Which capabilities can this dataset support?
It can support training models to combine information across reports, track changes in event status, organize events chronologically, and answer questions using source evidence. It can also be used to evaluate multi-hop question answering and timeline reconstruction.
How can I access or use the data?
Follow the download method, file format, and usage instructions provided on the product page. Before use, review the field structure and create training, validation, and test splits as appropriate for your task.
What should I consider about licensing and news content?
The applicable permissions are defined by the product page or accompanying license. Source reports may be subject to third-party copyright, so users should verify relevant rights and comply with applicable copyright, privacy, and news content usage requirements.

Can't find the data you need?

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

Get this Dataset

High-quality data, ready for your next AI breakthrough.

Cite this Work

@dataset{Mobiusi2026,
  title={Breaking News Event Timeline Reconstruction Question-Answering Dataset},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/03754bdade60e45589f68ab73d818b7f?dataset_scene_id=17},
  urldate={2026-10-26},
  keywords={breaking news dataset,news timeline reconstruction,multi-hop question answering,event reasoning,news event tracking,news understanding training data,evidence-grounded QA,reasoning SFT dataset},
  version={1.0}
}

Using this in research? Please cite us.

placeholder
placeholder
placeholder
placeholder
placeholder
placeholder
placeholder

Popular Dataset Searches