When Do Pre-Training Biases Propagate to Downstream Tasks? A Case Study in Text Summarization
Faisal Ladhak, Esin Durmus, Mirac Suzgun, Tianyi Zhang, Dan Jurafsky, Kathleen McKeown, Tatsunori Hashimoto
Abstract
Large language models (LLMs) are subject to sociocultural and other biases previously identified using intrinsic evaluations. However, when and how these intrinsic biases in pre-trained LM representations propagate to downstream, fine-tuned NLP tasks like summarization is not well understood. In this work, we investigate one type of bias—name-nationality bias—and trace it from the pre-training stage to a downstream summarization task across multiple summarization modeling choices. We show that these biases manifest themselves as hallucinations in summarization, leading to factually incorrect summaries. We also find that this propagation of biases is algorithm-dependent: more abstractive models allow biases to propagate more directly to downstream tasks as hallucinated facts. Building on these observations, we further analyze how changes to the adaptation method and fine-tuning data set affect name nationality biases and show that while they can reduce the overall rate of hallucinations, they do not change the types of biases that do appear.- Anthology ID:
- 2023.eacl-main.234
- Volume:
- Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics
- Month:
- May
- Year:
- 2023
- Address:
- Dubrovnik, Croatia
- Editors:
- Andreas Vlachos, Isabelle Augenstein
- Venue:
- EACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 3206–3219
- Language:
- URL:
- https://aclanthology.org/2023.eacl-main.234
- DOI:
- 10.18653/v1/2023.eacl-main.234
- Cite (ACL):
- Faisal Ladhak, Esin Durmus, Mirac Suzgun, Tianyi Zhang, Dan Jurafsky, Kathleen McKeown, and Tatsunori Hashimoto. 2023. When Do Pre-Training Biases Propagate to Downstream Tasks? A Case Study in Text Summarization. In Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics, pages 3206–3219, Dubrovnik, Croatia. Association for Computational Linguistics.
- Cite (Informal):
- When Do Pre-Training Biases Propagate to Downstream Tasks? A Case Study in Text Summarization (Ladhak et al., EACL 2023)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-2/2023.eacl-main.234.pdf