PolERo: Studying Political Evasion in Romanian
Summary
The paper introduces PolERo, a dataset of 3,574 annotated question-answer pairs from Romanian presidential transcripts, to study political evasion. It evaluates various NLP classification models and examines cross-lingual transfer, finding that fine-tuned encoders are competitive and ambivalent evasion categories are challenging.
View Cached Full Text
Cached at: 09/03/26, 05:53 AM
# PolERo: Studying Political Evasion in Romanian Source: [https://arxiv.org/abs/2609.02391](https://arxiv.org/abs/2609.02391) [View PDF](https://arxiv.org/pdf/2609.02391) > Abstract:Political evasion refers to responses that engage with a question while withholding the requested information\. Recent NLP work frames political evasion as a classification task using a two\-level taxonomy of response clarity and fine\-grained evasion strategies\. Existing work on response clarity and evasion classification is limited to English, leaving open whether the taxonomy and model behavior transfer across languages and political contexts\. We introduce PolERo, a dataset of 3,574 human\-annotated question\-answer pairs extracted from official transcripts of five Romanian presidents\. We evaluate multiple classification approaches on both datasets under matched conditions, including TF\-IDF baselines, fine\-tuned encoder models, a proposed sliding\-window encoder, and zero/few\-shot LLM prompting\. We study cross\-lingual transfer through joint bilingual training and machine\-translation\-based data augmentation\. Our results indicate that fine\-tuned encoders are competitive, cross\-lingual transfer is asymmetric, and ambivalent evasion categories involving pragmatic cues remain the main challenge across all model families\. ## Submission history From: Gabriel Ştefan \[[view email](https://arxiv.org/show-email/346da67b/2609.02391)\] **\[v1\]**Wed, 2 Sep 2026 10:02:57 UTC \(4,703 KB\)
Similar Articles
Duluth at SemEval-2026 Task 6: DeBERTa with LLM-Augmented Data for Unmasking Political Question Evasions
University of Minnesota Duluth team used DeBERTa-V3-base augmented with synthetic data from Gemini 3 and Claude Sonnet 4.5 to classify political question evasions, achieving 8th place at SemEval-2026 Task 6.
Polar: A Benchmark for Evaluating Political Bias in LLMs
Polar is a 4,026-instance multiple-choice benchmark for evaluating political bias in LLMs across U.S. and South Korean political contexts, measuring bias through option-level likelihoods. Experiments on 38 LLMs show systematic bias patterns varying by political context, issue category, and presentation language.
Cross-lingual Relation Extraction with Large Language Models: Zero-Shot, Few-Shot, and Fine-Tuned Evaluation on Romanian
This paper investigates cross-lingual relation extraction for Romanian by translating the SemEval-2010 Task 8 benchmark and evaluating Gemma 4 under zero-shot, few-shot, and QLoRA fine-tuning, comparing with smaller encoder baselines.
Framing the Narrative: Ideological Mimicry in Large Language Models
The paper introduces the Poli-SHIFT dataset and evaluation framework showing that LLMs exhibit "ideological mimicry," systematically shifting political stance toward signals in user prompts, with terminology changes alone reversing model positions in 16.9% of matched comparisons across seven open-weight models.
A Corpus of Persuasion Techniques in Slavic Languages
A new annotated corpus of persuasion techniques in Bulgarian, Polish, and Russian, covering parliamentary debates and social media, with 25 fine-grained techniques and baseline models for detection and classification.