README.md

---
license: other
license_name: research-use-only
license_link: LICENSE
language:
- ar
- de
- en
- es
- hi
- pl
- pt
- ru
- uk
multilinguality:
- multilingual
task_categories:
- text-classification
tags:
- propaganda-detection
- disinformation
- stance-detection
- search-engines
- multilingual
- russian-propaganda
pretty_name: 'SWARM: Search-Web documents Annotated for Russian propaganda, Multilingual'
size_categories:
- 1K<n<10K
extra_gated_heading: Request access to SWARM
extra_gated_description: >-
  Access is reviewed manually by the authors. Requests without a clear research
  purpose or an identifiable affiliation will not be approved.
extra_gated_prompt: >-
  SWARM is released for non-commercial research use only. The documents originate
  from public search engine results and may be subject to the source engines'
  terms of service and to third-party copyright. By requesting access you agree to
  use the data solely for research, not to redistribute it, and to honour takedown
  requests. Full terms are in the LICENSE file in this repository.
extra_gated_fields:
  Full name: text
  Institutional email: text
  Affiliation (university, institute, or company): text
  Country: country
  Role:
    type: select
    options:
      - Undergraduate or master's student
      - PhD student
      - Postdoc or research staff
      - Faculty
      - Industry researcher
      - Journalist
      - label: Other
        value: other
  How will you use SWARM? Describe the research question and the analysis or models you intend to run: text
  Do you intend to publish results derived from this dataset?:
    type: select
    options:
      - 'Yes, in a peer-reviewed venue'
      - 'Yes, as a preprint or report'
      - 'No, internal research only'
      - Undecided
  I will use this dataset for non-commercial research only: checkbox
  I will not redistribute this dataset or any substantial portion of it, and will refer others to this repository: checkbox
  I will honour takedown requests and delete affected records on notification: checkbox
  I will not use this dataset to generate, amplify, or target propaganda, or to profile individual authors or publishers: checkbox
  I agree to cite the accompanying paper in any publication that uses this dataset: checkbox
extra_gated_button_content: Request access
configs:
- config_name: default
  data_files:
  - split: test
    path: swarm.parquet
dataset_info:
  features:
  - name: id
    dtype: int64
  - name: lang
    dtype: string
  - name: narrative_question
    dtype: string
  - name: label
    dtype: int64
  - name: text_en
    dtype: string
  - name: text_native
    dtype: string
  - name: content_hash
    dtype: string
  splits:
  - name: test
    num_examples: 2129
---

# SWARM

**SWARM** (**S**earch-**W**eb documents **A**nnotated for **R**ussian propaganda, **M**ultilingual) is a multilingual, human-annotated dataset for detecting support for Russian propaganda narratives in **search engine results**.

It contains **2,129 documents** returned by four major search engines in response to queries about twenty recurring pro-Kremlin narratives, across **nine languages** (Arabic, English, German, Hindi, Polish, Portuguese, Russian, Spanish, Ukrainian). Each document was annotated by trained human coders for whether it **supports** a given narrative, yielding narrative-specific labels rather than broad propaganda-presence judgments.

> **Content warning.** This dataset contains propaganda narratives, some with dehumanising elements, reproduced only to document and study the phenomenon.

## Task

Given a **document** returned by a search engine and a **narrative** phrased as a single yes/no **question**, decide whether the document *supports* the narrative (answer to the question is *yes*). A document counts as supportive when it openly endorses the narrative (presenting it as fact or amplifying it uncritically), not merely when it mentions the topic. The negative class groups outright denial together with neutral or unrelated coverage, so the operative distinction is **support versus everything else**.

This is related to stance detection but harder: the target is a diffuse narrative rather than a sharply bounded proposition, and a document can support a narrative without asserting any single proposition. It also differs from fact verification, whose target is the *veracity* of a claim: a document can advance a propaganda narrative while making claims that are literally true, unfalsifiable, or value-laden.

## Data fields

Single split (`test`), one row per document:

| field | type | description |
|---|---|---|
| `id` | int64 | Unique document identifier |
| `lang` | string | **Query** language (`ar`, `de`, `en`, `es`, `hi`, `pl`, `pt`, `ru`, `uk`) |
| `narrative_question` | string | The pro-Kremlin narrative operationalised as a yes/no question (English) |
| `label` | int64 | **1** if the document supports the narrative, **0** otherwise |
| `text_en` | string | Document text in English (original for `en`, machine-translated otherwise), truncated to 12,000 characters. Null for 10 documents (see Known limitations) |
| `text_native` | string | Document text as retrieved, truncated to 12,000 characters. Never null |
| `content_hash` | string | Hash of the document content (for deduplication / joining) |

Both `text_en` and `text_native` are provided so the benchmark can be run either on English-normalised input or on native-language input. The labels and questions are identical across the two.

**`lang` is the language of the search query, not of the document.** Sampling is stratified by the language in which the search engine was queried, and engines routinely return results in other languages: a Hindi-language query, for instance, frequently returns English pages. Do not read `lang` as a document-language label, and do not use it as ground truth for language identification.

## Statistics

- **2,129** documents, **384 positive (18.0%)**
- **25** distinct narrative-question phrasings covering **20** pro-Kremlin narratives, organised around five actor groups (the United States, "the West", the European Union, Russia, Ukraine)

| lang | n | positive | positive % |
|---|---|---|---|
| ar | 258 | 46 | 17.8 |
| de | 244 | 52 | 21.3 |
| en | 217 | 44 | 20.3 |
| es | 227 | 34 | 15.0 |
| hi | 210 | 17 | 8.1 |
| pl | 244 | 24 | 9.8 |
| pt | 239 | 41 | 17.2 |
| ru | 244 | 89 | 36.5 |
| uk | 246 | 37 | 15.0 |

## Usage

```python
from datasets import load_dataset

ds = load_dataset("manueltonneau/SWARM", split="test")
ex = ds[0]
print(ex["narrative_question"])   # e.g. "do sanctions against russia primarily harm ..."
print(ex["label"])                # 1 = supports the narrative
print(ex["text_en"][:500])
```

Report metrics **per language** rather than only in aggregate: search results are a snapshot of specific regions and a time window, and both models and coders carry language-dependent behaviour. Positive-class F1 on the minority (support) class is the headline metric.

## How it was built

Twenty narratives drawn from recurring themes in Russian state media were each turned into two search queries (one propaganda-framed, one neutral) and one yes/no question, translated into nine languages by native speakers. Search engines were audited with these queries; returned results were sampled, the linked pages/PDFs scraped, and each document annotated by trained coders for narrative support. Coding ran in two passes that differ in what the coder saw: a live-page pass, in which the coder opened the result URL and judged the page as rendered, and a text-based pass on the extracted English text. A document is included only if at least one text-based coder could assess it. The gold label is the majority of all usable votes, with ties dropped, and a further coder was brought in where the first two disagreed. Document text is truncated to 12,000 characters, matching the truncation used in the paper's benchmarks.

## Known limitations

The dataset is released with its residual extraction and annotation noise intact rather than silently filtered, so that the item count matches the benchmark reported in the paper. Users running their own evaluation may wish to exclude the affected records.

- **Failed translations (10 documents).** The translation pass that produced `text_en` occasionally returned a meta-response instead of a translation: 6 exceeded the output budget and came back as a request to translate the text in parts, 3 were refused as explicit content, and 1 could not be decoded from its source encoding. Those strings are not document content, so `text_en` is set to **null** for the 10 affected records. Their labels are unaffected and all 10 are label 0. For one of them the `text_native` is itself corrupted (mojibake). Filter with `ds.filter(lambda x: x["text_en"] is not None)` when evaluating on English input.
- **Extraction failures (roughly 15 documents).** A small number of records are bot-protection interstitials, "access denied" pages, or consent walls rather than article text, having passed the extraction filter because they contain a well-formed body of prose. They were annotated as presented. Most carry label 0, but three carry label 1.
- **Label conflicts (2 pairs).** Four records form two pairs with identical text and identical `narrative_question` but opposite labels. All four are bot-protection pages, so the coders were in effect annotating boilerplate rather than content. They are retained for transparency.
- **Near-duplicate text.** Nine groups of records (18 rows total) share a `content_hash`, mostly because the same page was returned by more than one engine or region. Deduplicate on `content_hash` if independence across items matters for your analysis.
- **Selection effect on the positive class.** Extraction does not succeed uniformly across sources. Russian state-linked domains are over-represented among the failures, and because such domains are more likely to carry propaganda-supporting content, the retained set may slightly under-count the positive class.

## Intended use and access

SWARM is intended for **non-commercial research** on propaganda detection, multilingual NLP, and search-engine auditing. Because the documents come from public search results and may be subject to source engines' terms of service and third-party copyright, the dataset is distributed under **gated, research-use-only access** with a request process and a takedown procedure. It is **not** intended for training or deploying systems that generate, amplify, or target propaganda.

## Citation

```bibtex
@misc{tonneau2026swarm,
  title  = {SWARM: A Multilingual Human-Annotated Dataset for Russian Propaganda Detection in Search Engine Results},
  author = {Tonneau, Manuel and Dubey, Abhinav and Shaikh, Farhan and Vitulano, Ilaria and Stolze, Martha and Dedeoglu, Hale and Riechert, Clara and Kuka, Ella and Sydorova, Maryna and Makhortykh, Mykola and Kuznetsova, Elizaveta},
  year   = {2026}
}
```