Skip to content
conftrace_
2025 ACL ACL 2025

Multilabel Classification of Persuasion Techniques with self-improving LLM agent: SlavicNLP 2025 Shared Task

Abstract

AbstractWe present a system for the SlavicNLP 2025 Shared Task on multilabel classification of 25 persuasion techniques across Slavic languages. We investigate the effectiveness of in-context learning with one-shot classification, automatic prompt refinement, and supervised fine-tuning using self-generated annotations. Our findings highlight the potential of LLM-based system to generalize across languages and label sets with minimal supervision.

🌉 Interdisciplinary Bridge - Artificial Intelligence and Natural Language Processing

Authors

Marcin Sawinski, Krzysztof Wecel, Ewelina Ksiezniak

Topics

Artificial Intelligence > Core AI > Agent Systems Artificial Intelligence > Learning Paradigms > Few-Shot Learning Machine Learning > Core Methods > Classification Natural Language Processing > Applications > Text Classification Artificial Intelligence > Core AI > Large Language Models Machine Learning > Learning Types > In-Context Learning Deep Learning > Models > Large Language Models Machine Learning > Learning Types > Multi-Label Classification

Keywords

in-context learning prompt engineering multi-label classification persuasion detection llm agent large language model self-improving agent self-improving system
Download PDF

More from ACL 2025

All ACL 2025 papers →
LLMs for Detection and Classification of Persuasion Techniques in Slavic Parliamentary Debates and Social Media Texts
Fine-Tuned Transformer-Based Weighted Soft Voting Ensemble for Persuasion Technique Classification in Slavic Languages
Robust Detection of Persuasion Techniques in Slavic Languages via Multitask Debiasing and Walking Embeddings
SlavicNLP 2025 Shared Task: Detection and Classification of Persuasion Techniques in Parliamentary Debates and Social Media
Enhancing Retrieval for ESGLLM via ESG-CID: A Disclosure Content Index Finetuning Dataset for Mapping GRI and ESRS
Judging It, Washing It: Scoring and Greenwashing Corporate Climate Disclosures using Large Language Models