# Open-DeBias: Toward Mitigating Open-Set Bias in Language Models

arXiv: https://arxiv.org/abs/2509.23805 · HTML: https://mlsanity.com/paper/2509.23805/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 27 of its 37 references that are in the map.
- First posted on arXiv 28 Sep 2025.

## Builds on (most cited first, 27 of 27)

- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding](https://mlsanity.com/paper/1804.07461/) (2018)
- [Parameter-Efficient Transfer Learning for NLP](https://mlsanity.com/paper/1902.00751/) (2019)
- [Unsupervised Cross-lingual Representation Learning at Scale](https://mlsanity.com/paper/1911.02116/) (2019)
- [Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context](https://mlsanity.com/paper/2403.05530/) (2024)
- [Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings](https://mlsanity.com/paper/1607.06520/) (2016)
- [DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing](https://mlsanity.com/paper/2111.09543/) (2021)
- [Semantics derived automatically from language corpora contain human-like biases](https://mlsanity.com/paper/1608.07187/) (2016)
- [RACE: Large-scale ReAding Comprehension Dataset From Examinations](https://mlsanity.com/paper/1704.04683/) (2017)
- [AdapterFusion: Non-Destructive Task Composition for Transfer Learning](https://mlsanity.com/paper/2005.00247/) (2020)
- [BBQ: A Hand-Built Bias Benchmark for Question Answering](https://mlsanity.com/paper/2110.08193/) (2021)
- [MAD-X: An Adapter-Based Framework for Multi-Task Cross-Lingual Transfer](https://mlsanity.com/paper/2005.00052/) (2020)
- [Examining Gender and Race Bias in Two Hundred Sentiment Analysis Systems](https://mlsanity.com/paper/1805.04508/) (2018)
- [UnQovering Stereotyping Biases via Underspecified Questions](https://mlsanity.com/paper/2010.02428/) (2020)
- [Sustainable Modular Debiasing of Language Models](https://mlsanity.com/paper/2109.03646/) (2021)
- [On Learning to Summarize with Large Language Models as References](https://mlsanity.com/paper/2305.14239/) (2023)
- [Mitigating Political Bias in Language Models Through Reinforced Calibration](https://mlsanity.com/paper/2104.14795/) (2021)
- [KoBBQ: Korean Bias Benchmark for Question Answering](https://mlsanity.com/paper/2307.16778/) (2023)
- [Ethical-Advice Taker: Do Language Models Understand Natural Language Interventions?](https://mlsanity.com/paper/2106.01465/) (2021)
- [Gender Bias in BERT -- Measuring and Analysing Biases through Sentiment Rating in a Realistic Downstream Classification Task](https://mlsanity.com/paper/2306.15298/) (2023)
- [On the Robustness of Language Encoders against Grammatical Errors](https://mlsanity.com/paper/2005.05683/) (2020)
- [Increasing the LLM Accuracy for Question Answering: Ontologies to the Rescue!](https://mlsanity.com/paper/2405.11706/) (2024)
- [Detecting Unintended Social Bias in Toxic Language Datasets](https://mlsanity.com/paper/2210.11762/) (2022)
- [Mitigating Bias for Question Answering Models by Tracking Bias Influence](https://mlsanity.com/paper/2310.08795/) (2023)
- [Toward Deconfounding the Influence of Entity Demographics for Question Answering Accuracy](https://mlsanity.com/paper/2104.07571/) (2021)
- [The Synthetic Imputation Approach: Generating Optimal Synthetic Texts For Underrepresented Categories In Supervised Classification Tasks](https://mlsanity.com/paper/2504.15160/) (2025)

## Abstract (arXiv)

Large Language Models (LLMs) have achieved remarkable success on question answering (QA) tasks, yet they often encode harmful biases that compromise fairness and trustworthiness. Most existing bias mitigation approaches are restricted to predefined categories, limiting their ability to address novel or context-specific emergent biases. To bridge this gap, we tackle the novel problem of open-set bias detection and mitigation in text-based QA. We introduce OpenBiasBench, a comprehensive benchmark designed to evaluate biases across a wide range of categories and subgroups, encompassing both known and previously unseen biases. Additionally, we propose Open-DeBias, a novel, data-efficient, and parameter-efficient debiasing method that leverages adapter modules to mitigate existing social and stereotypical biases while generalizing to unseen ones. Compared to the state-of-the-art BMBI method, Open-DeBias improves QA accuracy on BBQ dataset by nearly 48% on ambiguous subsets and 6% on disambiguated ones, using adapters fine-tuned on just a small fraction of the training data. Remarkably, the same adapters, in a zero-shot transfer to Korean BBQ, achieve 84% accuracy, demonstrating robust language-agnostic generalization. Through extensive evaluation, we also validate the effectiveness of Open-DeBias across a broad range of NLP tasks, including StereoSet and CrowS-Pairs, highlighting its robustness, multilingual strength, and suitability for general-purpose, open-domain bias mitigation. The project page is available at: https://sites.google.com/view/open-debias25