File size: 13,666 Bytes
b0ede6c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
---
language:
  - "en"
license: "cc-by-nc-4.0"
library_name: "transformers"
pipeline_tag: text-classification
tags:
  - "text"
  - "politics"
  - "political"
  - "leaning"
  - "bias"
  - "politicalness"
base_model: "microsoft/deberta-v3-large"
datasets:
  - "mlburnham/dem_rep_party_platform_topics"
  - "cajcodes/political-bias"
  - "JyotiNayak/political_ideologies"
  - "Jacobvs/PoliticalTweets"
widget:
  - example_title: "Taxes 1"
    text: "The government should raise taxes on the rich so it can give more money to the homeless."
    output:
      - label: left
        score: 1.00
      - label: center
        score: 0.00
      - label: right
        score: 0.00
  - example_title: "Taxes 2"
    text: "The government should cut taxes because it is not using them efficiently anyway."
    output:
      - label: left
        score: 0.00
      - label: center
        score: 0.00
      - label: right
        score: 1.00
  - example_title: "Abortion 1"
    text: "Opting for abortion is an inalienable right of every individual."
    output:
      - label: left
        score: 1.00
      - label: center
        score: 0.00
      - label: right
        score: 0.00
  - example_title: "Abortion 2"
    text: "Terminating a pregnancy is equivalent to committing homicide."
    output:
      - label: left
        score: 0.42
      - label: center
        score: 0.00
      - label: right
        score: 0.58
  - example_title: "Immigration 1"
    text: "Mass detention of undocumented persons is an unjust practice that disproportionately harms vulnerable populations and must end."
    output:
      - label: left
        score: 1.00
      - label: center
        score: 0.00
      - label: right
        score: 0.00
  - example_title: "Immigration 2"
    text: "Immigration must be strictly controlled to protect national security, as it increases the risk of terrorism."
    output:
      - label: left
        score: 0.00
      - label: center
        score: 0.00
      - label: right
        score: 1.00
model-index:
  - name: "political-leaning-deberta-large"
    results:
      - task:
          type: "text-classification"
          name: "text political leaning classification"
        dataset:
          type: "-"
          name: "Article bias prediction"
        metrics:
          - type: "f1"
            value: 89
            name: "F1 score"
            args:
              average: "weighted"
        source:
          name: "the paper"
          url: "https://github.com/matous-volf/political-leaning-prediction/blob/main/paper.pdf"
      - task:
          type: "text-classification"
          name: "text political leaning classification"
        dataset:
          type: "-"
          name: "BIGNEWSBLN"
        metrics:
          - type: "f1"
            value: 88.6
            name: "F1 score"
            args:
              average: "weighted"
        source:
          name: "the paper"
          url: "https://github.com/matous-volf/political-leaning-prediction/blob/main/paper.pdf"
      - task:
          type: "text-classification"
          name: "text political leaning classification"
        dataset:
          type: "-"
          name: "CommonCrawl news articles"
        metrics:
          - type: "f1"
            value: 88.9
            name: "F1 score"
            args:
              average: "weighted"
        source:
          name: "the paper"
          url: "https://github.com/matous-volf/political-leaning-prediction/blob/main/paper.pdf"
      - task:
          type: "text-classification"
          name: "text political leaning classification"
        dataset:
          type: "-"
          name: "Dem., rep. party platform topics"
        metrics:
          - type: "f1"
            value: 85.6
            name: "F1 score"
            args:
              average: "weighted"
        source:
          name: "the paper"
          url: "https://github.com/matous-volf/political-leaning-prediction/blob/main/paper.pdf"
      - task:
          type: "text-classification"
          name: "text political leaning classification"
        dataset:
          type: "cajcodes/political-bias"
          name: "GPT-4 political bias"
        metrics:
          - type: "f1"
            value: 86.9
            name: "F1 score"
            args:
              average: "weighted"
        source:
          name: "the paper"
          url: "https://github.com/matous-volf/political-leaning-prediction/blob/main/paper.pdf"
      - task:
          type: "text-classification"
          name: "text political leaning classification"
        dataset:
          type: "JyotiNayak/political_ideologies"
          name: "GPT-4 political ideologies"
        metrics:
          - type: "f1"
            value: 99.6
            name: "F1 score"
            args:
              average: "weighted"
        source:
          name: "the paper"
          url: "https://github.com/matous-volf/political-leaning-prediction/blob/main/paper.pdf"
      - task:
          type: "text-classification"
          name: "text political leaning classification"
        dataset:
          type: "-"
          name: "Media political stance"
        metrics:
          - type: "f1"
            value: 93.1
            name: "F1 score"
            args:
              average: "weighted"
        source:
          name: "the paper"
          url: "https://github.com/matous-volf/political-leaning-prediction/blob/main/paper.pdf"
      - task:
          type: "text-classification"
          name: "text political leaning classification"
        dataset:
          type: "-"
          name: "Political podcasts"
        metrics:
          - type: "f1"
            value: 99.8
            name: "F1 score"
            args:
              average: "weighted"
        source:
          name: "the paper"
          url: "https://github.com/matous-volf/political-leaning-prediction/blob/main/paper.pdf"
      - task:
          type: "text-classification"
          name: "text political leaning classification"
        dataset:
          type: "Jacobvs/PoliticalTweets"
          name: "Political tweets"
        metrics:
          - type: "f1"
            value: 82.1
            name: "F1 score"
            args:
              average: "weighted"
        source:
          name: "the paper"
          url: "https://github.com/matous-volf/political-leaning-prediction/blob/main/paper.pdf"
      - task:
          type: "text-classification"
          name: "text political leaning classification"
        dataset:
          type: "-"
          name: "Qbias"
        metrics:
          - type: "f1"
            value: 57.9
            name: "F1 score"
            args:
              average: "weighted"
        source:
          name: "the paper"
          url: "https://github.com/matous-volf/political-leaning-prediction/blob/main/paper.pdf"
---

# Text political leaning classifier based on DeBERTa V3 large

This model classifies text by its political leaning into three classes: left, center, right. It has been trained on news
articles, social network posts and LLM-generated politological statements. The training data comes from the context of
the United States, and so the left class is mostly defined by the liberal ideology and democratic party views, and the
same applies for the right class being closely tied to the conservative and republican views.

The model is a part of the research done in the paper
[Predicting political leaning and politicalness of text using transformer models](https://github.com/matous-volf/political-leaning-prediction/blob/main/paper.pdf).
It focuses on predicting political leaning as well as politicalness – a binary class indicating whether a text even is
about politics or not. We have benchmarked the existing models for politicalness and shown that one of them –
[Political DEBATE](https://huggingface.co/mlburnham/Political_DEBATE_large_v1.0) – achieves an \\(F_1\\) score of over
90 %. This makes it suitable for filtering non-political texts in front of a political leaning classifier like this
one. We recommend doing so if the input to this model is not guaranteed to be about politics.

Our paper addresses the challenge of automatically classifying text according to political leaning and politicalness
using transformer models. We compose a comprehensive overview of existing datasets and models for these tasks, finding
that current approaches create siloed solutions that perform poorly on out-of-distribution texts. To address this
limitation, we compile a diverse dataset by combining 12 datasets for political leaning classification and creating a
new dataset for politicalness by extending 18 existing datasets with the appropriate label. Through extensive
benchmarking with leave-one-in and leave-one-out methodologies, we evaluate the performance of existing models and train
new ones with enhanced generalization capabilities.

Alongside the paper, we release the complete
[source code and results](https://github.com/matous-volf/political-leaning-prediction). This model is deployed in
a [demo web app](https://political-leaning.matousvolf.cz).
A [second, smaller model](https://huggingface.co/matous-volf/political-leaning-politics) has also been produced.

## Usage

The model outputs 0 for the left, 1 for the center and 2 for the right leaning. The score of the predicted class is
between \\(\frac{1}{3}\\) and 1.

To use the model, you can either utilize the high-level Hugging Face
[pipeline](https://huggingface.co/docs/transformers/main_classes/pipelines):

```py
from transformers import pipeline

pipe = pipeline(
    "text-classification",
    model="matous-volf/political-leaning-deberta-large",
    tokenizer="microsoft/deberta-v3-large",
)

text = "The government should raise taxes on the rich so it can give more money to the homeless."

output = pipe(text)
print(output)
```

Or load it [directly](https://huggingface.co/docs/transformers/en/models):

```py
from torch import argmax
from transformers import AutoModelForSequenceClassification, AutoTokenizer
from torch.nn.functional import softmax

tokenizer = AutoTokenizer.from_pretrained("microsoft/deberta-v3-large")
model = AutoModelForSequenceClassification.from_pretrained("matous-volf/political-leaning-deberta-large")

text = "The government should cut taxes because it is not using them efficiently anyway."

tokens = tokenizer(text, return_tensors="pt")
output = model(**tokens)
logits = output.logits

political_leaning = argmax(logits, dim=1).item()
probabilities = softmax(logits, dim=1)
score = probabilities[0, political_leaning].item()
print(political_leaning, score)
```

## Evaluation

The following table displays the performance of the model on test sets (15 %) of the datasets used for training.

| dataset                          | accuracy | \\(F_1\\) score |
|:---------------------------------|:---------|:----------------|
| Article bias prediction          | 89       | 89              |
| BIGNEWSBLN                       | 88.6     | 88.6            |
| CommonCrawl news articles        | 88.9     | 88.9            |
| Dem., rep. party platform topics | 85.5     | 85.6            |
| GPT-4 political bias             | 87       | 86.9            |
| GPT-4 political ideologies       | 99.6     | 99.6            |
| Media political stance           | 91.6     | 93.1            |
| Political podcasts               | 99.8     | 99.8            |
| Political tweets                 | 82.1     | 82.1            |
| Qbias                            | 58       | 57.9            |
| **average**                      | **87**   | **87.2**        |

The following is an example of a confusion matrix, after evaluating the model on a test set from the CommonCrawl news
articles dataset.

<img src="confusion_matrix.svg" alt="a confusion matrix example" height="350rem"/>

The complete results of all our measurements are available in the source code repository.

## Training

This model is based on [DeBERTa V3 large](https://huggingface.co/microsoft/deberta-v3-large). All the datasets used for
fine-tuning are listed in the paper, as well as a detailed description of the preprocessing, training and evaluation
methodology. In summary, we have manually tweaked the hyperparameters with a setup designed for maximizing performance
on unseen types of text (out-of-distribution) to increase the model's generalization abilities. In this setup, we have
left one of the datasets at a time out of the training sample and used it as the validation set. Then, we have taken the
resulting optimal hyperparameters and trained this model on all the available datasets.

## Authors

- Matous Volf ([[email protected]](mailto:[email protected])),
  [DELTA – High school of computer science and economics](https://www.delta-skola.cz), Pardubice, Czechia
- Jakub Simko ([[email protected]](mailto:[email protected])),
  [Kempelen Institute of Intelligent Technologies](https://kinit.sk), Bratislava, Slovakia

## Citation

### BibTeX

```
@article{volf-simko-2025-political-leaning,
  title        = {Predicting political leaning and politicalness of text using transformer models},
  author       = {Volf, Matous and Simko, Jakub},
  year         = 2025,
  institution  = {DELTA – High school of computer science and economics, Pardubice, Czechia; Kempelen Institute of Intelligent Technologies, Bratislava, Slovakia}
}
```

### APA

Volf, M. and Simko, J. (2025). Predicting political leaning and politicalness of text using transformer models. DELTA –
High school of computer science and economics, Pardubice, Czechia; Kempelen Institute of Intelligent Technologies,
Bratislava, Slovakia.