Metraipolis › archive › dc_07102ab669f0
Causal Evidence Extraction and Triangulation in Crisis Reports using Large Language Models: A ReliefWeb-based Study — arXiv:2608.04576
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.04576
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T11:24:44.386Z
- SHA-256 of the archived text
8311374d378f72d72fb960d4cf8d4aef8067830da6b19b2921b286e248a8728c- Stable URL of this page
https://metraipolis.com/documento/dc_07102ab669f0
How to cite
Causal Evidence Extraction and Triangulation in Crisis Reports using Large Language Models: A ReliefWeb-based Study — arXiv:2608.04576 — arXiv (en), https://arxiv.org/abs/2608.04576 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_07102ab669f0 (Metraipolis archive dc_07102ab669f0, sha256 8311374d378f72d72fb960d4cf8d4aef8067830da6b19b2921b286e248a8728c)
BibTeX
@misc{dc07102ab669f0,
title = {Causal Evidence Extraction and Triangulation in Crisis Reports using Large Language Models: A ReliefWeb-based Study — arXiv:2608.04576},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.04576},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_07102ab669f0; sha256 8311374d378f72d72fb960d4cf8d4aef8067830da6b19b2921b286e248a8728c},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.04576. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Humanitarian reports are long, noisy, and multi-topic, making it difficult to consolidate decision-relevant causal evidence. We present a ReliefWeb study (2000-2024) and a two-stage Large Language Model (LLM) pipeline that extracts structured intervention-outcome records with direction and strength attributes. Query-conditioned extraction restricts output to a specified intervention class, reducing retrieval-induced over-extraction, while snippet grounding links each relation to supporting text for auditability and classification. In an expert-annotated dataset of 100 reports, the best closed-source LLM achieved a weighted F1 score of 90.73% with strong cost-efficiency, while Llama-3.1-8B with supervised fine-tuning reached 94.15% weighted F1 score. We further propose context-preserving triangulation that aggregates strength-weighted evidence within disaster$\times$source cells, applies Laplace smoothing and equally weights cells to quantify cross-context convergence via a Level-of-Evidence score. Applied to cash assistance, food-related outcomes show strong positive convergence (LoE=0.865) and stable long-horizon trajectories. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-06 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2608.04576 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-31T11:13:23.180Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-06?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).