Traditional ML bias mitigation methods outperform LLM-based methods in both fairness and predictive performance, with prior LLM advantages driven by artificially balanced test data rather than realistic imbalanced distributions.
InProceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
LLMs Are Not a Silver Bullet: A Case Study on Software Fairness
Traditional ML bias mitigation methods outperform LLM-based methods in both fairness and predictive performance, with prior LLM advantages driven by artificially balanced test data rather than realistic imbalanced distributions.