Language, Culture, and Ideology: Personalizing Offensiveness Detection in Political Tweets with Reasoning LLMs
Fuente:
arXiv
Salvato in:
| Autori principali: | , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866918153225764864 |
|---|---|
| author | Pihulski, Dzmitry Kocoń, Jan |
| author_facet | Pihulski, Dzmitry Kocoń, Jan |
| contents | We explore how large language models (LLMs) assess offensiveness in political discourse when prompted to adopt specific political and cultural perspectives. Using a multilingual subset of the MD-Agreement dataset centered on tweets from the 2020 US elections, we evaluate several recent LLMs - including DeepSeek-R1, o4-mini, GPT-4.1-mini, Qwen3, Gemma, and Mistral - tasked with judging tweets as offensive or non-offensive from the viewpoints of varied political personas (far-right, conservative, centrist, progressive) across English, Polish, and Russian contexts. Our results show that larger models with explicit reasoning abilities (e.g., DeepSeek-R1, o4-mini) are more consistent and sensitive to ideological and cultural variation, while smaller models often fail to capture subtle distinctions. We find that reasoning capabilities significantly improve both the personalization and interpretability of offensiveness judgments, suggesting that such mechanisms are key to adapting LLMs for nuanced sociopolitical text classification across languages and ideologies. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_02351 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Language, Culture, and Ideology: Personalizing Offensiveness Detection in Political Tweets with Reasoning LLMs Pihulski, Dzmitry Kocoń, Jan Computation and Language Artificial Intelligence We explore how large language models (LLMs) assess offensiveness in political discourse when prompted to adopt specific political and cultural perspectives. Using a multilingual subset of the MD-Agreement dataset centered on tweets from the 2020 US elections, we evaluate several recent LLMs - including DeepSeek-R1, o4-mini, GPT-4.1-mini, Qwen3, Gemma, and Mistral - tasked with judging tweets as offensive or non-offensive from the viewpoints of varied political personas (far-right, conservative, centrist, progressive) across English, Polish, and Russian contexts. Our results show that larger models with explicit reasoning abilities (e.g., DeepSeek-R1, o4-mini) are more consistent and sensitive to ideological and cultural variation, while smaller models often fail to capture subtle distinctions. We find that reasoning capabilities significantly improve both the personalization and interpretability of offensiveness judgments, suggesting that such mechanisms are key to adapting LLMs for nuanced sociopolitical text classification across languages and ideologies. |
| title | Language, Culture, and Ideology: Personalizing Offensiveness Detection in Political Tweets with Reasoning LLMs |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2510.02351 |