SocioEval: A Template-Based Framework for Evaluating Socioeconomic Status Bias in Foundation Models

Divyanshu Kumar; Ishita Gupta; Nitin Aravind Birur; Tanay Baswa; Sahil Agarwal; Prashanth Harshangi

cs.CLApr 3, 2026

SocioEval: A Template-Based Framework for Evaluating Socioeconomic Status Bias in Foundation Models

Authors: Divyanshu Kumar, Ishita Gupta, Nitin Aravind Birur, Tanay Baswa, Sahil Agarwal, Prashanth Harshangi

Organizations: Enkrypt AI

Abstract

As Large Language Models (LLMs) increasingly power decision-making systems across critical domains, understanding and mitigating their biases becomes essential for responsible AI deployment. Although bias assessment frameworks have proliferated for attributes such as race and gender, socioeconomic status bias remains significantly underexplored despite its widespread implications in the real world. We introduce SocioEval, a template-based framework for systematically evaluating socioeconomic bias in foundation models through decision-making tasks. Our hierarchical framework encompasses 8 themes and 18 topics, generating 240 prompts across 6 class-pair combinations. We evaluated 13 frontier LLMs on 3,120 responses using a rigorous three-stage annotation protocol, revealing substantial variation in bias rates (0.42%-33.75%). Our findings demonstrate that bias manifests differently across themes lifestyle judgments show 10 $\times$ higher bias than education-related decisions and that deployment safeguards effectively prevent explicit discrimination but show brittleness to domain-specific stereotypes. SocioEval provides a scalable, extensible foundation for auditing class-based bias in language models.

Figures & tables

Figure 1: Overview of the SocioEval framework. (Left) The hierarchical data structure comprises 8 comprehensive themes, 18 fine-grained topics, 4 distinct class identities, and 6 class pair combinations. (Right) A three-stage evaluation protocol applied to 3,120 model responses, consisting of contextual review (prompt-response analysis and reasoning patterns), classification (biased or unbiased), and quality assurance (validation and agreement checks).

Figure 2: Bias rates across 13 frontier LLMs. Anthropic models show lowest bias rates, while Mistral models exhibit highest. Error bars represent 95% confidence intervals.

Figure 3: Bias rates by theme and class pair. Lifestyle themes show highest bias, while education-related themes show lowest. Extreme class pairs (upper vs. working) elicit more bias than adjacent pairs.

Figure 4: Distribution of response strategies across models. Anthropic models show high refusal rates; open-source models show varied patterns. Class preference is the most common biased response type.

Appendix figures & tables1 asset

Supplementary material from the paper’s appendix.

Appendix

Figure 5: Distribution of fine-grained response classifications across 13 models. The six categories reveal distinct patterns: Anthropic models show high refusal rates, while models with high bias rates predominantly exhibit class preference. This visualization demonstrates the value of granular classification in understanding model behavior.

Social Bias in Language Models

Social bias in language models is the study and mitigation of unfair stereotypes and disparities in generated text, judgments, and decisions.

See all 199 papers

A Scalable Entity-Based Framework for Auditing Bias in Large Language Models
Jan 18, 2026Akram Elbouanani, Aboubacar Tuo, Adrian PopescuLLM Auditing Political Bias in Language Models
Université Paris-Saclay, CEA, List, F-91120, Palaiseau, France
Guardrail-Agnostic Societal Bias Evaluation in Large Vision-Language Models
Aug 30, 2026Yusuke Hirota, Michael Ross Boone, Arun George Zachariah +4Language Model Bias Evaluation VLM Evaluation
NVIDIA
Evaluating Second-Order Bias of LLMs Through Epistemic Entitlement
Jun 16, 2026Ramaravind Kommiya Mothilal, Terry Jingchen Zhang, Raiyan Ahmed +3LLM-as-a-Judge Social Bias in Language Models
University of Toronto · Vector Institute · EuroSafeAI +1

Jan 18, 2026J/K move · Enter open · S save

A Scalable Entity-Based Framework for Auditing Bias in Large Language Models

Akram Elbouanani, Aboubacar Tuo, Adrian Popescu

Université Paris-Saclay, CEA, List, F-91120, Palaiseau, France

LLM Auditing Political Bias in Language Models Cultural Bias in Language Models Language Model Bias Evaluation

Open paper AskPDF arXiv

Existing approaches to bias evaluation in large language models (LLMs) trade ecological validity for statistical control, relying either on artificial prompts that poorly reflect real-world use or on naturalistic tasks that lack scale and rigor. We introduce a scalable bias-auditing framework that uses named entities as controlled probes to measure systematic disparities in model behavior. Synthetic data enables us to construct diverse, controlled inputs, and we show that it reliably reproduces bias patterns observed in natural text, supporting its use for large-scale analysis. Using this framework, we conduct the largest bias audit to date, comprising 1.9 billion data points across multiple entity types, tasks, languages, models, and prompting strategies. We find consistent patterns: models penalize right-wing politicians and favor left-wing politicians, prefer Western and wealthier countries over the Global South, favor Western companies, and penalize firms in the defense and pharmaceutical sectors. While instruction tuning reduces bias, increasing model scale amplifies it, and prompting in Chinese or Russian does not mitigate Western-aligned preferences. These findings highlight the need for systematic bias auditing before deploying LLMs in high-stakes applications. Our framework is extensible to other domains and tasks, and we make it publicly available to support future work.

Abstract

Social Bias in Language Models

Explore similar work

A Scalable Entity-Based Framework for Auditing Bias in Large Language Models

Guardrail-Agnostic Societal Bias Evaluation in Large Vision-Language Models

Evaluating Second-Order Bias of LLMs Through Epistemic Entitlement