ImplicitBBQ: Benchmarking Implicit Bias in Large Language Models through Characteristic Based Cues

• URL: http://arxiv.org/abs/2604.01925v1
• Date: Thu, 02 Apr 2026 11:43:10 GMT
• Title: ImplicitBBQ: Benchmarking Implicit Bias in Large Language Models through Characteristic Based Cues
• Abstract summary: Implicit bias in ambiguous contexts is over six times higher than explicit bias in open weight models.<n>Even few-shot prompting, which reduces implicit bias by 84%, leaves caste bias at four times the level of any other dimension.<n>We release our code and dataset for model providers and researchers to benchmark potential mitigation techniques.
• Score: 8.68634179512072
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large Language Models increasingly suppress biased outputs when demographic identity is stated explicitly, yet may still exhibit implicit biases when identity is conveyed indirectly. Existing benchmarks use name based proxies to detect implicit biases, which carry weak associations with many social demographics and cannot extend to dimensions like age or socioeconomic status. We introduce ImplicitBBQ, a QA benchmark that evaluates implicit bias through characteristic based cues, culturally associated attributes that signal implicitly, across age, gender, region, religion, caste, and socioeconomic status. Evaluating 11 models, we find that implicit bias in ambiguous contexts is over six times higher than explicit bias in open weight models. Safety prompting and chain-of-thought reasoning fail to substantially close this gap; even few-shot prompting, which reduces implicit bias by 84%, leaves caste bias at four times the level of any other dimension. These findings indicate that current alignment and prompting strategies address the surface of bias evaluation while leaving culturally grounded stereotypic associations largely unresolved. We publicly release our code and dataset for model providers and researchers to benchmark potential mitigation techniques.

Related papers

• Auditing demographic bias in AI-based emergency police dispatch: a cross-lingual evaluation of eleven large language models [3.85700834433791]
  We introduce a cross-lingual audit framework to operationalize the Police Priority Dispatch System.<n>We find that demographic bias emerges systematically when incident severity is ambiguous.<n>Gender bias is substantially amplified in Mandarin Chinese, whereas race bias is more pronounced in English.
  arXiv  Detail & Related papers  (2026-05-02T14:00:57Z)
• A Comprehensive Study of Implicit and Explicit Biases in Large Language Models [1.0555164678638427]
  This study highlights the need to address biases in Large Language Models amid growing generative AI.<n>We studied bias-specific benchmarks such as StereoSet and CrowSPairs to evaluate the existence of various biases in multiple generative models such as BERT and GPT 3.5.<n>Results indicated fine-tuned models struggle with gender biases but excelled at identifying and avoiding racial biases.
  arXiv  Detail & Related papers  (2025-11-18T05:27:17Z)
• Breaking the Benchmark: Revealing LLM Bias via Minimal Contextual Augmentation [12.56588481992456]
  Large Language Models have been shown to demonstrate stereotypical biases in their representations and behavior.<n>We introduce a novel and general augmentation framework that involves three plug-and-play steps.<n>We find that Large Language Models are susceptible to perturbations to their inputs, showcasing a higher likelihood to behave stereotypically.
  arXiv  Detail & Related papers  (2025-10-27T23:05:12Z)
• What's Not Said Still Hurts: A Description-Based Evaluation Framework for Measuring Social Bias in LLMs [8.219247185418821]
  Large Language Models (LLMs) often exhibit social biases inherited from their training data.<n>We introduce the Description-based Bias Benchmark (DBB), a novel dataset designed to assess bias at the semantic level.<n>We analyze six state-of-the-art LLMs, revealing that while models reduce bias in response at the term level, they continue to reinforce biases in nuanced settings.
  arXiv  Detail & Related papers  (2025-02-27T04:25:54Z)
• Covert Bias: The Severity of Social Views' Unalignment in Language Models Towards Implicit and Explicit Opinion [0.40964539027092917]
  We evaluate the severity of bias toward a view by using a biased model in edge cases of excessive bias scenarios. Our findings reveal a discrepancy in LLM performance in identifying implicit and explicit opinions, with a general tendency of bias toward explicit opinions of opposing stances. The direct, incautious responses of the unaligned models suggest a need for further refinement of decisiveness.
  arXiv  Detail & Related papers  (2024-08-15T15:23:00Z)
• Identifying and Mitigating Social Bias Knowledge in Language Models [52.52955281662332]
  We propose a novel debiasing approach, Fairness Stamp (FAST), which enables fine-grained calibration of individual social biases.<n>FAST surpasses state-of-the-art baselines with superior debiasing performance.<n>This highlights the potential of fine-grained debiasing strategies to achieve fairness in large language models.
  arXiv  Detail & Related papers  (2024-08-07T17:14:58Z)
• VLBiasBench: A Comprehensive Benchmark for Evaluating Bias in Large Vision-Language Model [72.13121434085116]
  We introduce VLBiasBench, a benchmark to evaluate biases in Large Vision-Language Models (LVLMs)<n>VLBiasBench features a dataset that covers nine distinct categories of social biases, including age, disability status, gender, nationality, physical appearance, race, religion, profession, social economic status, as well as two intersectional bias categories: race x gender and race x social economic status.<n>We conduct extensive evaluations on 15 open-source models as well as two advanced closed-source models, yielding new insights into the biases present in these models.
  arXiv  Detail & Related papers  (2024-06-20T10:56:59Z)
• SocialStigmaQA: A Benchmark to Uncover Stigma Amplification in Generative Language Models [8.211129045180636]
  We introduce a benchmark meant to capture the amplification of social bias, via stigmas, in generative language models. Our benchmark, SocialStigmaQA, contains roughly 10K prompts, with a variety of prompt styles, carefully constructed to test for both social bias and model robustness. We find that the proportion of socially biased output ranges from 45% to 59% across a variety of decoding strategies and prompting styles.
  arXiv  Detail & Related papers  (2023-12-12T18:27:44Z)
• GPTBIAS: A Comprehensive Framework for Evaluating Bias in Large Language Models [83.30078426829627]
  Large language models (LLMs) have gained popularity and are being widely adopted by a large user community. The existing evaluation methods have many constraints, and their results exhibit a limited degree of interpretability. We propose a bias evaluation framework named GPTBIAS that leverages the high performance of LLMs to assess bias in models.
  arXiv  Detail & Related papers  (2023-12-11T12:02:14Z)
• Social Bias Probing: Fairness Benchmarking for Language Models [38.180696489079985]
  This paper proposes a novel framework for probing language models for social biases by assessing disparate treatment. We curate SoFa, a large-scale benchmark designed to address the limitations of existing fairness collections. We show that biases within language models are more nuanced than acknowledged, indicating a broader scope of encoded biases than previously recognized.
  arXiv  Detail & Related papers  (2023-11-15T16:35:59Z)
• CBBQ: A Chinese Bias Benchmark Dataset Curated with Human-AI Collaboration for Large Language Models [52.25049362267279]
  We present a Chinese Bias Benchmark dataset that consists of over 100K questions jointly constructed by human experts and generative language models. The testing instances in the dataset are automatically derived from 3K+ high-quality templates manually authored with stringent quality control. Extensive experiments demonstrate the effectiveness of the dataset in detecting model bias, with all 10 publicly available Chinese large language models exhibiting strong bias in certain categories.
  arXiv  Detail & Related papers  (2023-06-28T14:14:44Z)
• "I'm sorry to hear that": Finding New Biases in Language Models with a Holistic Descriptor Dataset [12.000335510088648]
  We present a new, more inclusive bias measurement dataset, HolisticBias, which includes nearly 600 descriptor terms across 13 different demographic axes. HolisticBias was assembled in a participatory process including experts and community members with lived experience of these terms. We demonstrate that HolisticBias is effective at measuring previously undetectable biases in token likelihoods from language models.
  arXiv  Detail & Related papers  (2022-05-18T20:37:25Z)
• The SAME score: Improved cosine based bias score for word embeddings [49.75878234192369]
  We introduce SAME, a novel bias score for semantic bias in embeddings. We show that SAME is capable of measuring semantic bias and identify potential causes for social bias in downstream tasks.
  arXiv  Detail & Related papers  (2022-03-28T09:28:13Z)
• Balancing out Bias: Achieving Fairness Through Training Reweighting [58.201275105195485]
  Bias in natural language processing arises from models learning characteristics of the author such as gender and race. Existing methods for mitigating and measuring bias do not directly account for correlations between author demographics and linguistic variables. This paper introduces a very simple but highly effective method for countering bias using instance reweighting.
  arXiv  Detail & Related papers  (2021-09-16T23:40:28Z)

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.