Abstract
This work adapts large language models to generate multilingual social media text that meets several objectives simultaneously: topic relevance, author style consistency, and reply validity. Leveraging existing online information behavior simulators, which currently only forecast activities but not content, our approach comprised of generalizable prompt formation and efficient evaluation to produce a believable, personalized, and responsive synthetic social network. According to some preliminary experiments, our multi-objective prompt formation and automatic evaluation/selection methods are able to yield a significant number of high-quality synthetic texts according to both standardized and trained metrics.- Anthology ID:
- 2022.gem-1.39
- Volume:
- Proceedings of the 2nd Workshop on Natural Language Generation, Evaluation, and Metrics (GEM)
- Month:
- December
- Year:
- 2022
- Address:
- Abu Dhabi, United Arab Emirates (Hybrid)
- Editors:
- Antoine Bosselut, Khyathi Chandu, Kaustubh Dhole, Varun Gangal, Sebastian Gehrmann, Yacine Jernite, Jekaterina Novikova, Laura Perez-Beltrachini
- Venue:
- GEM
- SIG:
- SIGGEN
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 417–427
- Language:
- URL:
- https://aclanthology.org/2022.gem-1.39
- DOI:
- 10.18653/v1/2022.gem-1.39
- Cite (ACL):
- Mack Blackburn. 2022. Multilingual Social Media Text Generation and Evaluation with Few-Shot Prompting. In Proceedings of the 2nd Workshop on Natural Language Generation, Evaluation, and Metrics (GEM), pages 417–427, Abu Dhabi, United Arab Emirates (Hybrid). Association for Computational Linguistics.
- Cite (Informal):
- Multilingual Social Media Text Generation and Evaluation with Few-Shot Prompting (Blackburn, GEM 2022)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-2/2022.gem-1.39.pdf