Debiasing Algorithm through Model Adaptation
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Keywords: Model Editingi, Gender Bias, Causal Inference
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
Abstract: Large language models are becoming the go-to solution for the ever-growing number of tasks.
However, with growing capacity, models are prone to rely on spurious correlations stemming from biases and stereotypes present in the training data.
This work proposes a novel method for detecting and mitigating gender bias in language models.
We perform causal analysis to identify problematic model components and discover that mid-upper feed-forward layers are most prone to convey bias.
Based on the analysis results, we intervene in the model by applying a linear projection to the weight matrices of these layers.
Our titular method DAMA, significantly decreases bias as measured by diverse metrics while maintaining the model's performance on downstream tasks.
We release code for our method and models, which retrain LLaMA's state-of-the-art performance while being significantly less biased.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Primary Area: societal considerations including fairness, safety, privacy
Submission Number: 1695
Loading