Citation Needed Detection (Multilingual Wikipedia)
Sentence-level Citation Needed Detection (CND): given a claim from a Wikipedia article (with surrounding context), decide whether it requires an inline citation to a reliable source. Based on the Multilingual Citation Needed (MCN) corpus (Quaremba et al., ACL 2026), which spans 18 languages across three resource levels. Following prior CND work (Redi et al., 2019), corpus labels are derived from the presence or absence of inline citations; this task reproduces the underlying human verifiability judgment as a binary decision.
Configuration Fileconfig.yaml
This Potato config reproduces the annotation task. Save it as config.yaml and run potato start config.yaml to try it.
# Citation Needed Detection (CND) — Multilingual Wikipedia
# Based on the Multilingual Citation Needed (MCN) corpus:
# Quaremba, Rechkemmer, Black, Vrandečić, and Simperl (2026),
# "Multilingual and Cross-Lingual Citation Needed Detection on Wikipedia
# for Lower-Resource Languages." ACL 2026 (Vol. 1: Long Papers), pp. 28001-28021.
# Paper: https://aclanthology.org/2026.acl-long.1291/
# Data/code announced at: https://github.com/gerritq/mcn
#
# Task: given a sentence-level CLAIM from a Wikipedia article, together with the
# surrounding paragraph context, decide whether the claim requires an inline
# citation to a reliable source. This is the well-established binary CND
# formulation (Redi et al., 2019); in the MCN corpus, gold labels are derived
# from the presence/absence of inline citations, and here a human annotator makes
# the underlying verifiability judgment directly.
annotation_task_name: "Citation Needed Detection"
task_dir: "."
data_files:
- sample-data.json
item_properties:
id_key: "id"
text_key: "claim"
output_annotation_dir: "annotation_output/"
output_annotation_format: "json"
port: 8000
server_name: localhost
annotation_schemes:
- annotation_type: radio
name: citation_needed
description: "Does this claim require an inline citation to a reliable source?"
labels:
- "Citation Needed"
- "No Citation Needed"
keyboard_shortcuts:
"Citation Needed": "1"
"No Citation Needed": "2"
tooltips:
"Citation Needed": "The claim states a fact/figure/quotation that a reader could reasonably want verified against a source"
"No Citation Needed": "Common knowledge, a definition, an internal summary, or otherwise not requiring a source"
annotation_instructions: |
You are shown a CLAIM taken from a Wikipedia article, along with the paragraph
it appears in for context. Decide whether the claim would require an inline
citation to a reliable source under Wikipedia's verifiability policy.
Mark "Citation Needed" for claims likely to be challenged: statistics, specific
dates or figures, quotations, contested or surprising assertions, and statements
attributing views to people or organizations. Mark "No Citation Needed" for
common knowledge, definitions, plot/internal summaries, or purely structural
sentences. Judge the claim itself; the surrounding context is only to help you
understand it.
html_layout: |
<div style="padding: 15px; max-width: 820px; margin: auto;">
<div style="background: #f1f5f9; border: 1px solid #cbd5e1; border-radius: 8px; padding: 6px 12px; margin-bottom: 8px; font-size: 13px; color: #475569;">
<strong>Article:</strong> {{title}} · <strong>Language:</strong> {{language}}
</div>
<div style="background: #fffbeb; border: 1px solid #fde68a; border-radius: 8px; padding: 16px; margin-bottom: 8px;">
<strong style="color: #92400e;">Claim:</strong>
<p style="font-size: 17px; line-height: 1.6; margin: 8px 0 0 0;">{{claim}}</p>
</div>
<div style="color: #6b7280; font-size: 14px; line-height: 1.6;">
<strong>Paragraph context:</strong> {{context}}
</div>
</div>
allow_all_users: true
instances_per_annotator: 50
annotation_per_instance: 3
allow_skip: true
Sample Datasample-data.json
[
{
"id": "cnd_001",
"title": "Mount Kilimanjaro",
"language": "English",
"claim": "Mount Kilimanjaro rises approximately 4,900 metres from its base to a summit elevation of 5,895 metres above sea level.",
"context": "Mount Kilimanjaro is a dormant volcano in Tanzania. Mount Kilimanjaro rises approximately 4,900 metres from its base to a summit elevation of 5,895 metres above sea level. It is the highest mountain in Africa and the highest single free-standing mountain above sea level in the world."
},
{
"id": "cnd_002",
"title": "Photosynthesis",
"language": "English",
"claim": "Photosynthesis is a process used by plants and other organisms to convert light energy into chemical energy.",
"context": "Photosynthesis is a process used by plants and other organisms to convert light energy into chemical energy. This chemical energy is stored in carbohydrate molecules that are synthesized from carbon dioxide and water."
}
]
// ... and 8 more itemsGet This Design
Clone or download from the repository
Quick start:
git clone https://github.com/davidjurgens/potato-showcase.git cd potato-showcase/text/fact-verification/citation-needed-detection potato start config.yaml
Dataset & paper
Gerrit Quaremba, Amy Rechkemmer, Elizabeth Black, Denny Vrandečić, and Elena Simperl. 2026. Multilingual and Cross-Lingual Citation Needed Detection on Wikipedia for Lower-Resource Languages. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).
Citation (BibTeX)
@inproceedings{quaremba-etal-2026-multilingual,
title = "Multilingual and Cross-Lingual Citation Needed Detection on {W}ikipedia for Lower-Resource Languages",
author = "Quaremba, Gerrit and Rechkemmer, Amy and Black, Elizabeth and Vrande{\v{c}}i{\'c}, Denny and Simperl, Elena",
booktitle = "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
year = "2026",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-long.1291",
pages = "28001--28021"
}Details
Annotation Types
Domain
Use Cases
Tags
Found an issue or want to improve this design?
Open an IssueRelated Designs
FEVEROUS Structured Fact Verification Annotation
FEVEROUS is a fact-checking dataset of 87,026 Wikipedia claims with evidence drawn from both sentences and table cells. This Potato config reproduces the verdict, evidence, and justification task.
Table-Based Fact Verification (TabFact)
Verify textual claims against structured tabular data from Wikipedia. Based on TabFact (Chen et al., ICLR 2020). Annotators determine whether a natural language statement is entailed or refuted by the contents of a given table, highlight evidence cells, and provide reasoning.
Fact-Checking in Community Question Answering Forums
Verification of factual claims in community question answering forums, classifying answers to factual questions as true, false, or non-factual. Based on SemEval-2019 Task 8 (Fact-Checking in CQA), Subtask B.