Geographical Erasure in Language Generation
Pola Schwöbel, Jacek Golebiowski, Michele Donini, Cedric Archambeau, Danish Pruthi
Abstract
Large language models (LLMs) encode vast amounts of world knowledge. However, since these models are trained on large swaths of internet data, they are at risk of inordinately capturing information about dominant groups. This imbalance can propagate into generated language. In this work, we study and operationalise a form of geographical erasure wherein language models underpredict certain countries. We demonstrate consistent instances of erasure across a range of LLMs. We discover that erasure strongly correlates with low frequencies of country mentions in the training corpus. Lastly, we mitigate erasure by finetuning using a custom objective.- Anthology ID:
- 2023.findings-emnlp.823
- Volume:
- Findings of the Association for Computational Linguistics: EMNLP 2023
- Month:
- December
- Year:
- 2023
- Address:
- Singapore
- Editors:
- Houda Bouamor, Juan Pino, Kalika Bali
- Venue:
- Findings
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 12310–12324
- Language:
- URL:
- https://aclanthology.org/2023.findings-emnlp.823
- DOI:
- 10.18653/v1/2023.findings-emnlp.823
- Cite (ACL):
- Pola Schwöbel, Jacek Golebiowski, Michele Donini, Cedric Archambeau, and Danish Pruthi. 2023. Geographical Erasure in Language Generation. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 12310–12324, Singapore. Association for Computational Linguistics.
- Cite (Informal):
- Geographical Erasure in Language Generation (Schwöbel et al., Findings 2023)
- PDF:
- https://preview.aclanthology.org/alta-23-ingestion/2023.findings-emnlp.823.pdf