Pipelines for Social Bias Testing of Large Language Models
By Debora Nozza, Federico Bianchi, Dirk Hovy. 2022
The maturity level of language models is now at a stage in which many companies rely on them to solve various tasks. However, while research has shown how biased and harmful these models are, systematic ways of integrating social bias tests into development pipelines are still lacking. This short paper suggests how to use these verification techniques in development pipelines. We take inspiration from software testing and suggest addressing social bias evaluation as software testing. We hope to open a discussion on the best methodologies to handle social bias testing in language models.
Subfields: Large Language Models, Safety, Robustness & Interpretability
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Large Language Models
- Irene Solaiman
- Dirk Hovy
- Zeerak Talat
- Yacine Jernite
- Anne Lauscher
- Alexandra Sasha Luccioni
- Jesse Dodge
- Federico Bianchi
In Safety, Robustness & Interpretability
- Dirk Hovy
- Federico Bianchi
- Anne Lauscher
- Su Lin Blodgett
- Debora Nozza
- Irene Solaiman
- Zeerak Talat
- Dylan Baker
Citations in the corpus
- Evaluating the Social Impact of Generative AI Systems in Systems and Society (2023, background): “Base generative AI systems have no specific predefined application, sector, or use case, making them notoriously difficult to evaluate [166].”
- Gender bias and stereotypes in Large Language Models (2023, background): “However, as is well known, language models perpetuate and occasionally amplify biases, stereotypes, and negative perceptions of minoritized groups in society [10, 13, 14, 66, 69, 84, 85, 90].”
- Measuring Harmful Sentence Completion in Language Models for LGBTQIA+ Individuals (2022, methodology): “We believe that this contribution can be integrated into pipelines for the automatic evaluation of LLMs as described in (Nozza et al., 2022).”