# Invisible Influences: Investigating Implicit Intersectional Biases through Persona Engineering in Large Language Models Source: [Invisible Influences: Investigating Implicit Intersectional Biases through Persona Engineering in Large Language Models]() Collection index: [llms.txt]() Data updated: 2026-04-10T05:00:04.898Z Authors: Nandini Arimanda, Achyuth Mukund, Sakthi Balan Muthiah, Rajesh Sharma Published: 2026-03-16T15:57:12.000Z ## Abstract Large Language Models (LLMs) excel at human-like language generation but often embed and amplify implicit, intersectional biases, especially under persona-driven contexts. Existing bias audits rely on static, embedding-based tests (CEAT, I-WEAT, I-SEAT) that quantify absolute association strengths. We show that they have limitations in capturing dynamic shifts when models adopt social roles. We address this gap by introducing the Bias Amplification Differential and Explainability Score (BADx): a novel, scalable metric that measures persona-induced bias amplification and integrates local explainability insights. BADx comprises three components - differential bias scores (BAD, based on CEAT, I-WEAT, I-SEAT),Persona Sensitivity Index (PSI), and Volatility (Standard Deviation), augmented by LIME-based analysis for emphasizing explainability. This study is divided and performed as two different tasks. Task 1 establishes static bias baselines, and Task 2 applies six persona frames (marginalized and structurally advantaged) to measure BADx, PSI, and volatility. This is studied across five state-of-the-art LLMs (GPT-4o, DeepSeek-R1, LLaMA-4, Claude 4.0 Sonnet and Gemma-3n E4B). Results show persona context significantly modulates bias. GPT-4o exhibits high sensitivity and volatility; DeepSeek-R1 suppresses bias but with erratic volatility; LLaMA-4 maintains low volatility and a stable bias profile with limited amplification; Claude 4.0 Sonnet achieves balanced modulation; and Gemma-3n E4B attains the lowest volatility with moderate amplification. BADx performs better than static methods by revealing context-sensitive biases overlooked in static methods. Our unified method offers a systematic way to detect dynamic implicit intersectional bias in five popular LLMs. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Bias Evaluation on Task 2 Persona A](): Bias Evaluation; Task 2 Persona A; BAD Score: not recorded. - [Bias Evaluation on Task 2 Persona B](): Bias Evaluation; Task 2 Persona B; BAD Score: not recorded. - [Bias Evaluation on Task 2 Persona C](): Bias Evaluation; Task 2 Persona C; BAD Score: not recorded. - [Bias Evaluation on Task 2 Persona D](): Bias Evaluation; Task 2 Persona D; BAD Score: not recorded. - [Bias Evaluation on Task 2 Persona E](): Bias Evaluation; Task 2 Persona E; BAD Score: not recorded. - [Bias Evaluation on Task 2 Persona F](): Bias Evaluation; Task 2 Persona F; BAD Score: not recorded. - [Persona Bias Evaluation on Persona Engineering Persona A](): Persona Bias Evaluation; Persona Engineering Persona A; PSI: not recorded. - [Persona Bias Evaluation on Persona Engineering Persona B](): Persona Bias Evaluation; Persona Engineering Persona B; PSI: not recorded. - [Persona Bias Evaluation on Persona Engineering Persona C](): Persona Bias Evaluation; Persona Engineering Persona C; PSI: not recorded. - [Persona Bias Evaluation on Persona Engineering (Persona D)](): Persona Bias Evaluation; Persona Engineering (Persona D); PSI: not recorded. - [Persona Bias Evaluation on Persona Engineering Persona E](): Persona Bias Evaluation; Persona Engineering Persona E; PSI: not recorded. - [Persona Bias Evaluation on Persona Engineering Persona F](): Persona Bias Evaluation; Persona Engineering Persona F; PSI: not recorded. ## Original sources and implementation [Original paper]() [PDF]()