Newly appearing links concentrated around hubs in three news narratives
Source check: Source checked. Not independent scientific confirmation.
Finding: In language-model maps of coverage of three news events, 44% to 51% of next-day emerging links touched an entity in the top tenth by network degree. A ranking based on endpoint degree beat a historical-link baseline on these emerging links, and the margin shrank when high-degree entities were removed.
Technology: Claude Haiku, GPT-4.1 and Gemini semantic edge extraction
Family Claude: profile · models page
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Language models resolved references and extracted semantic relations from each article, which supplied the networks used to quantify where next-day links appeared.
- Test and result
- The retrospective comparison covered the Canadian Freedom Convoy, Hurricane Melissa and the 2025–26 Iranian protests. It used rolling windows and three separate language-model extractors.
- Limitation / still unknown
- The graphs use model-extracted relations without human gold labels. Candidate negatives were sampled randomly, favoring degree-based rankings. The result describes links represented in coverage of three events.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://arxiv.org/abs/2610.35628v1 · https://arxiv.org/html/2610.35628v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 21 Nov 2026. Review scope: §§3–5, Figures 8–9 and front matter.
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.