# Towards Understanding Task-agnostic Debiasing Through the Lenses of Intrinsic Bias and Forgetfulness

arXiv: https://arxiv.org/abs/2406.04146 · HTML: https://mlsanity.com/paper/2406.04146/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 30 of its 48 references that are in the map.
- First posted on arXiv 6 Jun 2024.

## Builds on (most cited first, 30 of 30)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Overcoming catastrophic forgetting in neural networks](https://mlsanity.com/paper/1612.00796/) (2016)
- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [A large annotated corpus for learning natural language inference](https://mlsanity.com/paper/1508.05326/) (2015)
- [SemEval-2017 Task 1: Semantic Textual Similarity - Multilingual and Cross-lingual Focused Evaluation](https://mlsanity.com/paper/1708.00055/) (2017)
- [StereoSet: Measuring stereotypical bias in pretrained language models](https://mlsanity.com/paper/2004.09456/) (2020)
- [Learning Sparse Neural Networks through L0 Regularization](https://mlsanity.com/paper/1712.01312/) (2017)
- [Null It Out: Guarding Protected Attributes by Iterative Nullspace Projection](https://mlsanity.com/paper/2004.07667/) (2020)
- [Bias in Bios: A Case Study of Semantic Representation Bias in a High-Stakes Setting](https://mlsanity.com/paper/1901.09451/) (2019)
- [Towards Understanding and Mitigating Social Biases in Language Models](https://mlsanity.com/paper/2106.13219/) (2021)
- [Mitigating Gender Bias in Natural Language Processing: Literature Review](https://mlsanity.com/paper/1906.08976/) (2019)
- [Variational Dropout Sparsifies Deep Neural Networks](https://mlsanity.com/paper/1701.05369/) (2017)
- [Measuring and Reducing Gendered Correlations in Pre-trained Models](https://mlsanity.com/paper/2010.06032/) (2020)
- [Identifying and Reducing Gender Bias in Word-Level Language Models](https://mlsanity.com/paper/1904.03035/) (2019)
- [Gender Bias in Neural Natural Language Processing](https://mlsanity.com/paper/1807.11714/) (2018)
- [Learning Gender-Neutral Word Embeddings](https://mlsanity.com/paper/1809.01496/) (2018)
- [A Fast Post-Training Pruning Framework for Transformers](https://mlsanity.com/paper/2204.09656/) (2022)
- [Intrinsic Bias Metrics Do Not Correlate with Application Bias](https://mlsanity.com/paper/2012.15859/) (2020)
- [Mechanistically analyzing the effects of fine-tuning on procedurally defined tasks](https://mlsanity.com/paper/2311.12786/) (2023)
- [Entropy-based Attention Regularization Frees Unintended Bias Mitigation from Lists](https://mlsanity.com/paper/2203.09192/) (2022)
- [Debiasing isn't enough! -- On the Effectiveness of Debiasing MLMs and their Social Biases in Downstream Tasks](https://mlsanity.com/paper/2210.02938/) (2022)
- [Double Descent Demystified: Identifying, Interpreting & Ablating the Sources of a Deep Learning Puzzle](https://mlsanity.com/paper/2303.14151/) (2023)
- [PAC-Bayes with Backprop](https://mlsanity.com/paper/1908.07380/) (2019)
- [Unified Detoxifying and Debiasing in Language Generation via Inference-time Adaptive Optimization](https://mlsanity.com/paper/2210.04492/) (2022)
- [MABEL: Attenuating Gender Bias using Textual Entailment Data](https://mlsanity.com/paper/2210.14975/) (2022)
- [Should We Attend More or Less? Modulating Attention for Fairness](https://mlsanity.com/paper/2305.13088/) (2023)
- [PAC-tuning:Fine-tuning Pretrained Language Models with PAC-driven Perturbed Gradient Descent](https://mlsanity.com/paper/2310.17588/) (2023)
- [On the Origins of Bias in NLP through the Lens of the Jim Code](https://mlsanity.com/paper/2305.09281/) (2023)

## Built on by (most cited first, 2 of 2)

- [Self-correction is Not An Innate Capability in Language Models](https://mlsanity.com/paper/2410.20513/) (2024)
- [Diagnosing the Performance Trade-off in Moral Alignment: A Case Study on Gender Stereotypes](https://mlsanity.com/paper/2509.21456/) (2025)

## Abstract (arXiv)

While task-agnostic debiasing provides notable generalizability and reduced reliance on downstream data, its impact on language modeling ability and the risk of relearning social biases from downstream task-specific data remain as the two most significant challenges when debiasing Pretrained Language Models (PLMs). The impact on language modeling ability can be alleviated given a high-quality and long-contextualized debiasing corpus, but there remains a deficiency in understanding the specifics of relearning biases. We empirically ascertain that the effectiveness of task-agnostic debiasing hinges on the quantitative bias level of both the task-specific data used for downstream applications and the debiased model. We empirically show that the lower bound of the bias level of the downstream fine-tuned model can be approximated by the bias level of the debiased model, in most practical cases. To gain more in-depth understanding about how the parameters of PLMs change during fine-tuning due to the forgetting issue of PLMs, we propose a novel framework which can Propagate Socially-fair Debiasing to Downstream Fine-tuning, ProSocialTuning. Our proposed framework can push the fine-tuned model to approach the bias lower bound during downstream fine-tuning, indicating that the ineffectiveness of debiasing can be alleviated by overcoming the forgetting issue through regularizing successfully debiased attention heads based on the PLMs' bias levels from stages of pretraining and debiasing.