Opinion events and stance types: advances in LLM performance with ChatGPT and Gemini
-
Barbara Lewandowska-Tomaszczyk
Barbara Lewandowska-Tomaszczyk is Professor Ordinarius Dr Habil. in Linguistics and English Language at the Department of Language and Communication at the University of Applied Sciences in Konin (Poland). Her research focuses on cognitive semantics and pragmatics of language contrasts, corpus linguistics and their applications in translation studies, lexicography and online discourse analysis. She is invited to read papers at international conferences and to lecture and conduct seminars at universities. She publishes extensively, supervises dissertations and organizes international conferences and workshops.
und
Chaya Liebeskind
Chaya Liebeskind is a lecturer and researcher in the Department of Computer Science at the Jerusalem College of Technology. Her research interests span both Natural Language Processing and Data Mining. Especially, her scientific interests include Semantic Similarity, Language Technology for Cultural Heritage, Morphologically Rich Languages (MRL), Multi-word Expressions (MWEs), Information Retrieval (IR), and Text Classification (TC). Much of her recent work has been focusing on analysing offensive language. She has published a variety of studies and a few of her articles are under review or in preparation. She is a member of several international research actions funded by the EU.
Abstract
The paper tests conversational Large Language Models, instructed to produce stance expression types (affective, relational, epistemic, and moral) and their contexts in Opinion (Speech) Events (Lewandowska-Tomaszczyk, Barbara, Chaya Liebeskind, Anna Baczkowska, Jurate Ruzaite, Ardita Dylgjeri, Ledia Kazazi & Erika Lombart 2023. Opinion events: Types and opinion markers in English social media discourse. Lodz Papers in Pragmatics 19(2). 447–481). In the first part an opinion taxonomy proposed in (Lewandowska-Tomaszczyk, Barbara, Chaya Liebeskind, Anna Baczkowska, Jurate Ruzaite, Ardita Dylgjeri, Ledia Kazazi & Erika Lombart 2023. Opinion events: Types and opinion markers in English social media discourse. Lodz Papers in Pragmatics 19(2). 447–481) is discussed in terms of Explicit (direct or indirect) and Implicit opinionated texts, categorized as positive, negative, ambiguous, or balanced. The further part discusses our previous attempts at Explicit (direct/indirect) and Implicit opinion type generation, performed by means of a series of prompts with LLMs (ChatGPT and Gemini) (Liebeskind, Chaya & Barbara Lewandowska-Tomaszczyk. 2024a. Opinion identification using a conversational large language model. In FLAIRS conference Proceedings. Florida, Liebeskind, Chaya & Barbara Lewandowska-Tomaszczyk. F 2024b. Navigating opinion space: A Study of explicit and implicit opinion generation in language models. Santiago de Compostella: EAIS conference publication), while this paper presents further LLM experiments with chatGPT and Gemini as well as their results, based on the analysis of stance expression types, which lead to increased success in opinion context generation.
About the authors
Barbara Lewandowska-Tomaszczyk is Professor Ordinarius Dr Habil. in Linguistics and English Language at the Department of Language and Communication at the University of Applied Sciences in Konin (Poland). Her research focuses on cognitive semantics and pragmatics of language contrasts, corpus linguistics and their applications in translation studies, lexicography and online discourse analysis. She is invited to read papers at international conferences and to lecture and conduct seminars at universities. She publishes extensively, supervises dissertations and organizes international conferences and workshops.
Chaya Liebeskind is a lecturer and researcher in the Department of Computer Science at the Jerusalem College of Technology. Her research interests span both Natural Language Processing and Data Mining. Especially, her scientific interests include Semantic Similarity, Language Technology for Cultural Heritage, Morphologically Rich Languages (MRL), Multi-word Expressions (MWEs), Information Retrieval (IR), and Text Classification (TC). Much of her recent work has been focusing on analysing offensive language. She has published a variety of studies and a few of her articles are under review or in preparation. She is a member of several international research actions funded by the EU.
Acknowledgments
The study was prepared in the Linguistics group of COST Action CA 21129 What are Opinions? Integrating Theory and Methods for Automatically Analyzing Opinionated Communication (OPINION).
References
Addawood, Aseel, Jodi Schneider & Masooda N. Bashir. 2017. Stance classification of Twitter debates: The encryption debate as A use case. #SMSociety17: Proceedings of the 8th International Conference on Social Media & Society. Article No.: 2.1–10. https://doi.org/10.1145/3097286.30972.Suche in Google Scholar
Aijmer, Karin. 1997. I think – an English modal particle. In T. Swan & O. Jansen Westvik (eds.), Modality in Germanic languages: Historical and comparative approaches, 1–47. Berlin: Mouton de Gruyter.10.1515/9783110889932.1Suche in Google Scholar
Anand, Pranav, Marilyn Walker, Rob Abbott, Jean E. Fox Tree, Robeson Bowmani & Michael Minor. 2011. Cats rule and dogs drool!: Classifying stance in online debate. In Alexandra Balahur, Ester Boldrini, Andres Montoyo & Patricio Martinez-Barco (eds.), Proceedings of the 2nd Workshop on Computational Approaches to Subjectivity and Sentiment Analysis (WASSA 2.011), 1–9. Portland, Oregon.Suche in Google Scholar
Du Bois, John W. 2007. The stance triangle. Stancetaking in Discourse: Subjectivity, Evaluation, Interaction 164(3). 139–182. https://doi.org/10.1075/pbns.164.07du.Suche in Google Scholar
Gray, Bethany & Douglas Biber. 2015. Stance markers. In Karin Aijmer & Christoph Rühlemann (eds.), Corpus pragmatics: A handbook, 219–248. Cambridge: Cambridge University Press.10.1017/CBO9781139057493.012Suche in Google Scholar
Hidalgo Downing, Raquel A. & Silvia Iglesias-Recuero. 2020. Humor e ironía, una relación compleja. In Leonor Ruiz Gurillo & Xose A. Padilla (eds.), Dime cómo ironizas y te diré quién eres. Una aproximación pragmática a la ironía, 423–457. Frankfurt a. Main: Peter Lang.Suche in Google Scholar
Hyland, Ken. 1999. Disciplinary discourses: Writer stance in research articles. In Christopher Candlin & Ken Hyland (eds.), Writing: Texts, processes and practices, 99–121. London, UK: Longman.10.4324/9781315840390-6Suche in Google Scholar
Ku, Lun-Wei, Li-Ying Lee & Hsin-Hsi Chen. 2009. Identification of opinion holders. International Journal of Computational Linguistics & Chinese Language Processing 14(4). 383–402.Suche in Google Scholar
Lewandowska-Tomaszczyk, Barbara, Chaya Liebeskind, Anna Baczkowska, Jurate Ruzaite, Ardita Dylgjeri, Ledia Kazazi & Erika Lombart. 2023. Opinion events: Types and opinion markers in English social media discourse. Lodz Papers in Pragmatics 19(2). 447–481. https://doi.org/10.1515/lpp-2023-0022.Suche in Google Scholar
Lewandowska-Tomaszczyk, Barbara, Olga, Dontcheva-Navrátilová & Renata, Povolná. (In press). Persuasion by emotion and stance expression in social media. In Olga, Dontcheva-Navrátilová & Renata, Povolná (eds.), Beyond the surface of the text. Analytical perspectives on text analysis. London, USA: Palgrave Macmillan.Suche in Google Scholar
Lian, Xiaoli, Dan Huang, Xuefeng Li, Ziyan Zhao, Zhiqiang Fan & Min Li. 2023. Really vague? Automatically identify the potential false vagueness within the context of documents. Mathematics 11(10). 2334. https://doi.org/10.3390/math11102334.Suche in Google Scholar
Liebeskind, Chaya & Barbara Lewandowska-Tomaszczyk. 2024a. Opinion identification using a conversational large language model. In FLAIRS conference Proceedings. Florida.10.32473/flairs.37.1.135529Suche in Google Scholar
Liebeskind, Chaya & Barbara Lewandowska-Tomaszczyk. 2024b. Navigating opinion space: A Study of explicit and implicit opinion generation in language models. Santiago de Compostella: EAIS conference publication.Suche in Google Scholar
Liu, Yufeng & Dechao Li. 2022. The US-China battle over Coronavirus in the news media: Metaphor transfer as a representation of stance mediation. Discourse & Society 33(4). 456–477. https://doi.org/10.1177/09579265221088122.Suche in Google Scholar
Ochs, Elinor & Lisa Capps. 2001. Living narrative: Creating lives in everyday storytelling. Cambridge: Harvard University Press.10.4159/9780674041592Suche in Google Scholar
Wei, Jason, Xiezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc Le & Denny Zhou. 2023. Chain-of-thought prompting elicits reasoning in large language models. Advances in Neural Information Processing Systems 35. 24824–24837.Suche in Google Scholar
White, Jules, Quchen Fu, Sam Hays, Michael Sandborn, Carlos Olea, Henry Gilbert, Ashraf El-nashar, Jessy Spencer Smith & Douglas C. Schmidt. 2023. A prompt pattern catalog to enhance prompt engineering with chatgpt. arXiv preprint arXiv:2302.11382.Suche in Google Scholar
Zhou, Denny, Nathanael Schärli, Le Hou, Jason Wei, Nathan Scales, Xuezhi Wang, Dale Schuurmans, Claire Cui, Olivier Bousquet & Quoc V. Le. 2022. Least-to-most prompting enables complex reasoning in large language models. ICLR 2023. arXiv preprint arXiv:2205.10625.Suche in Google Scholar
© 2024 Walter de Gruyter GmbH, Berlin/Boston
Artikel in diesem Heft
- Frontmatter
- Editorial
- Introduction to the Special Issue: Pragmatics, digital content and opinions
- Research Articles
- Discursive news values analysis: the case of Liz Truss’ representation in the British press
- The case of romantic relationships: analysis of the use of metaphorical frames with ‘traditional family’ and related terms in political Telegram posts in three countries and three languages
- Expressing anger on Mexican X/Twitter: the case of Uber customer complaints
- Expressing negative opinions through metaphor and simile in popular music reviews
- Slur reclamation, irony, and resilience
- What is the authentic internet register before & after the Russian invasion in Ukraine? Polish and Czech YouTube comments from 2021–2023
- Application of natural language processing for the recognition of obesity-related topics in the discourses of Argentine Twitter users
- Opinion events and stance types: advances in LLM performance with ChatGPT and Gemini
- Classifying offensive language in Arabic: a novel taxonomy and dataset
- Implicit offensive language taxonomy
Artikel in diesem Heft
- Frontmatter
- Editorial
- Introduction to the Special Issue: Pragmatics, digital content and opinions
- Research Articles
- Discursive news values analysis: the case of Liz Truss’ representation in the British press
- The case of romantic relationships: analysis of the use of metaphorical frames with ‘traditional family’ and related terms in political Telegram posts in three countries and three languages
- Expressing anger on Mexican X/Twitter: the case of Uber customer complaints
- Expressing negative opinions through metaphor and simile in popular music reviews
- Slur reclamation, irony, and resilience
- What is the authentic internet register before & after the Russian invasion in Ukraine? Polish and Czech YouTube comments from 2021–2023
- Application of natural language processing for the recognition of obesity-related topics in the discourses of Argentine Twitter users
- Opinion events and stance types: advances in LLM performance with ChatGPT and Gemini
- Classifying offensive language in Arabic: a novel taxonomy and dataset
- Implicit offensive language taxonomy
