In-Depth Look at Word Filling Societal Bias Measures
Pikuliak, M., Beňová, I., Bachratý, V.
Many measures of societal bias in language models have been proposed in recent years. A popular approach is to use a set of word filling prompts to evaluate the behavior of the language models. In this work, we analyze the validity of two such measures – StereoSet and CrowS-Pairs. We show that these measures produce unexpected and illogical results when appropriate control group samples are constructed. Based on this, we believe that they are problematic and using them in the future should be reconsidered. We propose a way forward with an improved testing protocol. Finally, we also introduce a new gender bias dataset for Slovak.
Cite: Pikuliak, M., Beňová, I., Bachratý, V. In-Depth Look at Word Filling Societal Bias Measures. In Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics, pages 3648–3665, Dubrovnik, Croatia. Association for Computational Linguistics (2023).