Proceedings of the MULTINOVA: First International Conference on Artificial Intelligence in Engineering, Healthcare and Sciences (ICAIEHS- 2025)

<Previous Article In Volume
Next Article In Volume>

Storytelling with Transformers: A Comparative Analysis of GPT-2 and BART

Authors
Hrithik Singh1, *, Vanita Mane1, Tushar Ghorpade1
1D Y Patil Deemed to be University, Nerul, India
*Corresponding author. Email: hri.sin.rt23@dypatil.edu
Corresponding Author
Hrithik Singh
Available Online 7 October 2025.
DOI
10.2991/978-94-6463-852-3_9How to use a DOI?
Keywords
Story generation; GPT-2; BART Transformers; NLP; Narrative evaluation
Abstract

This study presents a comparative analysis of two widely used transformer-based models, GPT-2 and BART, in the context of interactive storytelling. The evaluation emphasizes model performance in generating coherent, creative, and contextually aligned narratives. We utilize a Reddit-based dataset of writing prompts and apply fine-tuning techniques to both models. The experiment emphasizes the better performance of GPT-2 to generate semantically good and creative stories whereas BART generated a more emotionally consistent and grammatically perfect story. Those findings support the architectural differences of the models GPT-2 is autoregressive and thus was biased towards generating imaginatively, whereas BART has an encoder-decoder structure, which is biased towards controlled generation. Qualitative analysis showed that GPT-2 would tend to repeat similar words when given vague prompts and BART would tend to follow a narrative when given strange prompts. Our findings highlight each model’s strengths and limitations, offering insights into their capabilities for creative text generation and guiding future work in transformer-based storytelling systems.

Copyright
© 2025 The Author(s)
Open Access
Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

<Previous Article In Volume
Next Article In Volume>
Volume Title
Proceedings of the MULTINOVA: First International Conference on Artificial Intelligence in Engineering, Healthcare and Sciences (ICAIEHS- 2025)
Series
Advances in Intelligent Systems Research
Publication Date
7 October 2025
ISBN
978-94-6463-852-3
ISSN
1951-6851
DOI
10.2991/978-94-6463-852-3_9How to use a DOI?
Copyright
© 2025 The Author(s)
Open Access
Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Cite this article

risenwbib
TY  - CONF
AU  - Hrithik Singh
AU  - Vanita Mane
AU  - Tushar Ghorpade
PY  - 2025
DA  - 2025/10/07
TI  - Storytelling with Transformers: A Comparative Analysis of GPT-2 and BART
BT  - Proceedings of the MULTINOVA: First International Conference on Artificial Intelligence in Engineering, Healthcare and Sciences (ICAIEHS- 2025)
PB  - Atlantis Press
SP  - 139
EP  - 148
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6463-852-3_9
DO  - 10.2991/978-94-6463-852-3_9
ID  - Singh2025
ER  -