File size: 12,927 Bytes
0de825d 191e704 0de825d 191e704 0de825d 191e704 0de825d 191e704 a92697a bfc019a 0de825d e891d0c 8b3bae1 e891d0c 1e1df9c e891d0c 5a52fdf 365e5ab 5a52fdf 365e5ab e891d0c 365e5ab e891d0c 365e5ab e891d0c bfc019a e891d0c a92697a e891d0c 2f00ce9 e891d0c 2f00ce9 e891d0c 2f00ce9 e891d0c 6602080 c2e58be 6602080 e891d0c a92697a e891d0c 7fe89b8 a92697a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 |
---
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
dataset_info:
features:
- name: text
dtype: string
- name: labels
dtype: int64
splits:
- name: train
num_bytes: 1287816.4583728428
num_examples: 4217
- name: validation
num_bytes: 161244.2708135786
num_examples: 528
- name: test
num_bytes: 161244.2708135786
num_examples: 528
download_size: 965791
dataset_size: 1610304.9999999998
license: apache-2.0
language:
- es
pretty_name: ConspiraText-ES
size_categories:
- 1K<n<10K
---
# Dataset Card for "ConspiraText-ES"
---
annotations_creators:
- machine-generated
language_creators:
- machine-generated
languages:
- es
licenses:
- Apache 2.0
multilinguality:
- monolingual
pretty_name: ConspiraText-ES
size_categories:
- 1K<n<10K
source_datasets:
- original
task_categories:
- text-classification
task_ids:
- binary-classification
---
# Dataset Card for ConspiraText-ES
## Table of Contents
- [Dataset Description](#dataset-description)
- [Dataset Structure](#dataset-structure)
- [Dataset Creation](#dataset-creation)
- [Considerations for Using the Data](#considerations-for-using-the-data)
- [Additional Information](#additional-information)
## Dataset Description
**Homepage:** https://huggingface.co/MorcuendeA
**Repository:** https://huggingface.co/datasets/MorcuendeA/ConspiraText-ES
**Paper:** Not available
**Leaderboard:** Not available
**Point of Contact:** Not available
### Dataset Summary
ConspiraText-ES is a synthetic Spanish-language dataset, using DeepSeeker R1 0528, containing short, tweet-style messages related to 15 common conspiracy topics paired with their rational refutations. Each message is labeled as either conspiratorial or rational.
The dataset is balanced and designed for binary classification tasks.
### Supported Tasks and Leaderboards
- Text classification (binary): Detect conspiratorial vs. rational statements.
### Languages
- Spanish (`es`)
## Dataset Structure
### Data Instances
Each data instance consists of a short Spanish text message (tweet style) and a binary label indicating whether the message is conspiratorial (1) or rational (0).
text (string): A short Spanish text message.
labels (int): Class label where 0 = rational, 1 = conspiratorial.
### Data Splits
| Split | Total Samples | Class 0 | Class 1 |
|------------|---------------|---------|---------|
| Train | 4,217 | 2,096 | 2,121 |
| Validation | 528 | 256 | 272 |
| Test | 528 | 236 | 292 |
## Dataset Creation
### Curation Rationale
The dataset was created as a personal, experimental synthetic resource to simulate conspiratorial and rational discourse styles on key social topics. It aims to facilitate research in misinformation, discourse classification, and bias detection.
Financial Elites
Conspiracy: They control governments through debt.
Refutation: Complex and transparent economic processes.
Climate Change
Conspiracy: An excuse for a global government.
Refutation: A real and urgent scientific phenomenon.
Processed Foods
Conspiracy: Designed to reduce fertility.
Refutation: Regulated food safety.
Education System
Conspiracy: A tool for indoctrination.
Refutation: Formation of critical citizens.
Mass Media
Conspiracy: Fabricating pro-war consensus.
Refutation: Plurality of competing voices.
Space Exploration
Conspiracy: A hoax to divert funds.
Refutation: Science that generates innovation and collaboration.
Artificial Intelligence
Conspiracy: Elites use it to replace humans.
Refutation: A tool to complement human capabilities.
Vaccines
Conspiracy: A global surveillance program.
Refutation: Pillars of public health.
Pandemics
Conspiracy: A tool for social control.
Refutation: Real crises that require coordination.
Chemtrails
Conspiracy: They alter climate and human behavior.
Refutation: Harmless condensation trails.
Extraterrestrials
Conspiracy: Hidden by governments.
Refutation: No evidence; scientific skepticism.
5G Technology
Conspiracy: A mind control weapon.
Refutation: Safe telecommunications technology.
Social Networks
Conspiracy: Monitoring and shaping opinions.
Refutation: Spaces for free expression and information.
Large Fires
Conspiracy: Redesigning maps for corporate interests.
Refutation: Natural and human causes (climate change, negligence).
Official History
Conspiracy: Hides advanced civilizations.
Refutation: Academic consensus based on archaeological evidence.
### Source Data
This is an original dataset generated automatically by prompting the DeepSeeker R1 0528 language model.
### Initial Data Collection and Normalization
Text samples were generated using carefully designed prompts for 15 pairs of conspiracy topics and their rational counterarguments.
These prompts were randomly selected during dataset creation to ensure diversity and balance.
Below are examples of the prompts used to generate the two classes:
#### Conspiratorial prompts (Class 1):
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo las élites financieras controlan los gobiernos mediante deuda.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo el cambio climático es una excusa para imponer un gobierno global.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo los alimentos procesados están diseñados para reducir la fertilidad.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo el sistema educativo está hecho para adoctrinar y no para enseñar.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo los medios de comunicación fabrican consensos a favor de la guerra.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo la exploración espacial es una farsa para desviar fondos públicos.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo la inteligencia artificial está diseñada para reemplazar a los humanos por orden de las élites.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo las vacunas forman parte de un programa de vigilancia global.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo los gobiernos usan las pandemias como herramienta de control social.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo los chemtrails alteran el comportamiento humano y afectan el clima.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo los extraterrestres ya están entre nosotros pero los gobiernos lo ocultan.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo la tecnología 5G puede usarse como arma de manipulación mental.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo las redes sociales están diseñadas para vigilar y moldear opiniones.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo los grandes incendios son usados para rediseñar mapas urbanos según intereses corporativos.",
"Escríbeme un mensaje tipo tweet conspiranoico sobre cómo la historia oficial ha sido manipulada para esconder civilizaciones avanzadas del pasado."
#### Rational/refuting prompts (Class 0):
"Escríbeme un mensaje tipo tweet que explique por qué las élites financieras no controlan los gobiernos mediante deuda, sino a través de procesos económicos complejos y transparentes.",
"Escríbeme un mensaje tipo tweet que refute la idea de que el cambio climático es una excusa para un gobierno global, y explique por qué es un fenómeno científico real y urgente.",
"Escríbeme un mensaje tipo tweet que desmienta que los alimentos procesados están diseñados para reducir la fertilidad, y defienda la seguridad alimentaria regulada por agencias internacionales.",
"Escríbeme un mensaje tipo tweet que argumente que el sistema educativo no busca adoctrinar, sino formar ciudadanos críticos y preparados para el mundo real.",
"Escríbeme un mensaje tipo tweet que explique que los medios de comunicación no fabrican consensos a favor de la guerra, sino que existen múltiples voces y líneas editoriales en competencia.",
"Escríbeme un mensaje tipo tweet que defienda que la exploración espacial no es una farsa, sino un campo científico real que genera innovación, conocimiento y colaboración internacional.",
"Escríbeme un mensaje tipo tweet que critique la idea de que la inteligencia artificial está diseñada por las élites para reemplazar a los humanos, y explique cómo puede complementar nuestras capacidades.",
"Escríbeme un mensaje tipo tweet que desmienta que las vacunas formen parte de un programa de vigilancia global, y reafirme su importancia para la salud pública.",
"Escríbeme un mensaje tipo tweet que explique por qué las pandemias no son herramientas de control social, sino crisis reales que requieren coordinación global para ser contenidas.",
"Escríbeme un mensaje tipo tweet que refute la idea de que los chemtrails alteran el comportamiento humano o el clima, y explique qué son en realidad las estelas de los aviones.",
"Escríbeme un mensaje tipo tweet que cuestione la idea de que los extraterrestres ya están entre nosotros y que los gobiernos lo ocultan, y promueva el escepticismo científico.",
"Escríbeme un mensaje tipo tweet que critique la teoría de que la tecnología 5G puede usarse como arma de manipulación mental, y explique cómo funciona realmente esta tecnología.",
"Escríbeme un mensaje tipo tweet que defienda que las redes sociales no están diseñadas únicamente para vigilar o manipular opiniones, sino que también permiten la libre expresión y el acceso a información.",
"Escríbeme un mensaje tipo tweet que refute la idea de que los grandes incendios son parte de un plan para rediseñar mapas urbanos, y explique las causas reales detrás de estos desastres.",
"Escríbeme un mensaje tipo tweet que explique por qué la historia oficial no está manipulada para esconder civilizaciones avanzadas, y cómo funciona el consenso académico basado en evidencia."
#for rational
{"role": "system", "content": "Eres un generador de contenido que redacta mensajes tipo tweet con tono racional, escéptico y basado en hechos. Los mensajes deben parecer escritos por personas comunes que valoran el pensamiento crítico y la información verificada. Usa un lenguaje claro y accesible, sin tecnicismos excesivos. Genera cada tweet con una longitud media de 20-30 tokens. Es extremadamente importante que NO incluyas nada más en tu respuesta, escribe solamente el texto que te solicita sin ninguna aclaración o número de tokens."},
#for conspiratorial
{"role": "system", "content": "Eres un generador de contenido que redacta mensajes tipo tweet desde una perspectiva conspirativa. Los mensajes deben parecer escritos por alguien que cree sinceramente en lo que dice. Usa un tono serio, alarmista o revelador. El lenguaje debe ser informal, pero con convicción. Cubre cualquier tema conspiranoico (vacunas, política internacional, control social, aliens, etc.). Genera cada tweet con una longitud media de 20 tokens. No incluyas nada más en la salida, solo el texto del tweet."},
"temperature": 1,
"max_tokens": 512
### Who Are the Source Language Producers?
All text was machine-generated by the DeepSeeker R1 0528 model; no human-generated text is included.
### Annotations
#### Annotation Process
Labels were automatically assigned based on the prompt type: conspiratorial messages received label `1`; rational/refuting messages received label `0`.
#### Who Are the Annotators?
Annotations were generated automatically; no manual annotation was performed.
### Personal and Sensitive Information
No personal data is included. The dataset consists solely of synthetic, topic-based short texts.
## Considerations for Using the Data
### Social Impact of Dataset
This dataset can help develop tools to detect conspiracy theories and misinformation in Spanish text. However, care should be taken to avoid misuse or stigmatization of legitimate critical thinking.
### Discussion of Biases
- The dataset may contain biases originating from the language model used for generation.
- The binary label scheme simplifies complex social views.
- Stylistic biases related to prompt engineering may be present.
### Other Known Limitations
- Fully synthetic; does not reflect real human language or discourse nuance.
- Limited to 15 predefined topics.
- May not generalize to other languages or domains.
## Additional Information
### Dataset Curators
Created by [MorcuendeA].
### Licensing Information
This dataset is licensed under the Apache license 2.0.
### Citation Information
If you use this dataset, please cite as:
title = ConspiraText-ES: A Synthetic Spanish Dataset of Conspiratorial and Rational Statements
author = Morcuende A
year = 2025 |