PRIDE -- Parameter-Efficient Reduction of Identity Discrimination for Equality in LLMs

• URL: http://arxiv.org/abs/2507.13743v1
• Date: Fri, 18 Jul 2025 08:44:27 GMT
• Title: PRIDE -- Parameter-Efficient Reduction of Identity Discrimination for Equality in LLMs
• Abstract summary: Large Language Models (LLMs) frequently reproduce the gender- and sexual-identity prejudices embedded in their training corpora.<n>We evaluate two parameter-efficient fine-tuning (PEFT) techniques - Low-Rank Adaptation (LoRA) and soft-prompt tuning.<n>Fine-tuning with LoRA reduces those scores by up to 50 points and raises neutrality from virtually 0% to as much as 36%.
• Score: 0.0552480439325792
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large Language Models (LLMs) frequently reproduce the gender- and sexual-identity prejudices embedded in their training corpora, leading to outputs that marginalize LGBTQIA+ users. Hence, reducing such biases is of great importance. To achieve this, we evaluate two parameter-efficient fine-tuning (PEFT) techniques - Low-Rank Adaptation (LoRA) and soft-prompt tuning - as lightweight alternatives to full-model fine-tuning for mitigating such biases. Using the WinoQueer benchmark, we quantify bias in three open-source LLMs and observe baseline bias scores reaching up to 98 (out of 100) across a range of queer identities defined by gender and/or sexual orientation, where 50 would indicate neutrality. Fine-tuning with LoRA (< 0.1% additional parameters) on a curated QueerNews corpus reduces those scores by up to 50 points and raises neutrality from virtually 0% to as much as 36%. Soft-prompt tuning (10 virtual tokens) delivers only marginal improvements. These findings show that LoRA can deliver meaningful fairness gains with minimal computation. We advocate broader adoption of community-informed PEFT, the creation of larger queer-authored corpora, and richer evaluation suites beyond WinoQueer, coupled with ongoing audits to keep LLMs inclusive.

Related papers

• When More Parameters Hurt: Foundation Model Priors Amplify Worst-Client Disparity Under Extreme Federated Heterogeneity [0.0]
  Federated learning (FL) is increasingly used to fine-tune foundation models (FMs) on distributed private data.<n>We compare worst-client accuracy between TextCNN (2.7M parameters) and DistilBERT with Low-Rank Adaptation (LoRA)<n>Under extreme label skew (alpha = 0.1), DistilBERT+LoRA produces a worst-client accuracy gap of 50.1% -- 56% larger than TextCNN's 32.2% gap, despite having 25x more parameters and extensive pretraining.
  arXiv  Detail & Related papers  (2026-05-09T15:22:08Z)
• RefineBench: Evaluating Refinement Capability of Language Models via Checklists [71.02281792867531]
  We evaluate two refinement modes: guided refinement and self-refinement.<n>In guided refinement, both proprietary LMs and large open-weight LMs can leverage targeted feedback to refine responses to near-perfect levels within five turns.<n>These findings suggest that frontier LMs require breakthroughs to self-refine their incorrect responses.
  arXiv  Detail & Related papers  (2025-11-27T07:20:52Z)
• Beyond Consensus: Mitigating the Agreeableness Bias in LLM Judge Evaluations [0.20027036140258694]
  New Large Language Models (LLMs) become available every few weeks, and modern application developers confront with the unenviable task of deciding if they should switch to a new model.<n>We show that while LLMs can identify valid outputs with high accuracy, they are remarkably poor at identifying invalid ones.<n>We introduce an optimal minority-veto strategy that is resilient to missing data and mitigates this bias to a large extent.
  arXiv  Detail & Related papers  (2025-10-13T18:19:23Z)
• LLM Bias Detection and Mitigation through the Lens of Desired Distributions [8.710551702283391]
  We propose a weighted adaptive loss based fine-tuning method to align gender-profession output with desired distribution.<n>We achieve near-complete mitigation under equality and 30-75% reduction under real-world settings.
  arXiv  Detail & Related papers  (2025-10-07T18:20:03Z)
• Bias in Gender Bias Benchmarks: How Spurious Features Distort Evaluation [116.86965910589775]
  We show that even minimal perturbations, such as masking just 10% of objects or weakly blurring backgrounds, can dramatically alter bias scores.<n>This suggests that current bias evaluations reflect model responses to spurious features rather than gender bias.
  arXiv  Detail & Related papers  (2025-09-09T11:14:11Z)
• Towards Fair Rankings: Leveraging LLMs for Gender Bias Detection and Measurement [6.92803536773427]
  Social biases in Natural Language Processing (NLP) and Information Retrieval (IR) systems are an ongoing challenge.<n>We aim to address this issue by leveraging Large Language Models (LLMs) to detect and measure gender bias in passage ranking.<n>We introduce a novel gender fairness metric, named Class-wise Weighted Exposure (CWEx), aiming to address existing limitations.
  arXiv  Detail & Related papers  (2025-06-27T16:39:12Z)
• LFTF: Locating First and Then Fine-Tuning for Mitigating Gender Bias in Large Language Models [13.40656836132881]
  We propose datasets named GenBiasEval and GenHintEval.<n>The GenBiasEval is responsible for evaluating the degree of gender bias in LLMs.<n>The GenHintEval is used to assess whether LLMs can provide responses consistent with prompts that contain gender hints.
  arXiv  Detail & Related papers  (2025-05-21T12:49:37Z)
• The Root Shapes the Fruit: On the Persistence of Gender-Exclusive Harms in Aligned Language Models [91.86718720024825]
  We center transgender, nonbinary, and other gender-diverse identities to investigate how alignment procedures interact with pre-existing gender-diverse bias.<n>Our findings reveal that DPO-aligned models are particularly sensitive to supervised finetuning.<n>We conclude with recommendations tailored to DPO and broader alignment practices.
  arXiv  Detail & Related papers  (2024-11-06T06:50:50Z)
• Reward-Augmented Data Enhances Direct Preference Alignment of LLMs [63.32585910975191]
  We introduce reward-conditioned Large Language Models (LLMs) that learn from the entire spectrum of response quality within the dataset.<n>We show that our approach consistently boosts DPO by a considerable margin.<n>Our method not only maximizes the utility of preference data but also mitigates the issue of unlearning, demonstrating its broad effectiveness beyond mere data expansion.
  arXiv  Detail & Related papers  (2024-10-10T16:01:51Z)
• GenderCARE: A Comprehensive Framework for Assessing and Reducing Gender Bias in Large Language Models [73.23743278545321]
  Large language models (LLMs) have exhibited remarkable capabilities in natural language generation, but have also been observed to magnify societal biases.<n>GenderCARE is a comprehensive framework that encompasses innovative Criteria, bias Assessment, Reduction techniques, and Evaluation metrics.
  arXiv  Detail & Related papers  (2024-08-22T15:35:46Z)
• GenderBias-\emph{VL}: Benchmarking Gender Bias in Vision Language Models via Counterfactual Probing [72.0343083866144]
  This paper introduces the GenderBias-emphVL benchmark to evaluate occupation-related gender bias in Large Vision-Language Models. Using our benchmark, we extensively evaluate 15 commonly used open-source LVLMs and state-of-the-art commercial APIs. Our findings reveal widespread gender biases in existing LVLMs.
  arXiv  Detail & Related papers  (2024-06-30T05:55:15Z)
• GenderAlign: An Alignment Dataset for Mitigating Gender Bias in Large Language Models [20.98831667981121]
  Large Language Models (LLMs) are prone to generating content that exhibits gender biases.<n>GenderAlign dataset comprises 8k single-turn dialogues, each paired with a "chosen" and a "rejected" response.<n>Compared to the "rejected" responses, the "chosen" responses demonstrate lower levels of gender bias and higher quality.
  arXiv  Detail & Related papers  (2024-06-20T01:45:44Z)
• Disclosure and Mitigation of Gender Bias in LLMs [64.79319733514266]
  Large Language Models (LLMs) can generate biased responses. We propose an indirect probing framework based on conditional generation. We explore three distinct strategies to disclose explicit and implicit gender bias in LLMs.
  arXiv  Detail & Related papers  (2024-02-17T04:48:55Z)
• Marginal Debiased Network for Fair Visual Recognition [59.05212866862219]
  We propose a novel marginal debiased network (MDN) to learn debiased representations. Our MDN can achieve a remarkable performance on under-represented samples.
  arXiv  Detail & Related papers  (2024-01-04T08:57:09Z)
• Tokenization Matters: Navigating Data-Scarce Tokenization for Gender Inclusive Language Technologies [75.85462924188076]
  Gender-inclusive NLP research has documented the harmful limitations of gender binary-centric large language models (LLM) We find that misgendering is significantly influenced by Byte-Pair (BPE) tokenization. We propose two techniques: (1) pronoun tokenization parity, a method to enforce consistent tokenization across gendered pronouns, and (2) utilizing pre-existing LLM pronoun knowledge to improve neopronoun proficiency.
  arXiv  Detail & Related papers  (2023-12-19T01:28:46Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.