Downstream bias mitigation is all you need

Abstract: The advent of transformer-based architectures and LLMs have significantly advanced the performance of NLP models. Since these LLMs are trained on huge corpuses of data from the web and other sources, there has been a major concern about harmful prejudices that may potentially be transferred from the data. In many applications, these pre-trained LLMs are fine-tuned on task specific datasets, which can further contribute to biases. This paper studies the extent of biases absorbed by LLMs during pre-training as well as task-specific behaviour after fine-tuning. We found that controlled interventions on pre-trained LLMs, prior to fine-tuning, have minimal effect on lowering biases in classifiers. However, the biases present in domain-specific datasets play a much bigger role, and hence mitigating them at this stage has a bigger impact. While pre-training does matter, but after the model has been pre-trained, even slight changes to co-occurrence rates in the fine-tuning dataset has a significant effect on the bias of the model.