Article

Finding Rare Racial Stereotypes in LLM Responses via Adversarial Anomaly Detection

2026-05-18

Abstract excerpt

Large language models (LLMs) are increasingly aligned with antiracist norms through instruction tuning and reinforcement learning from human feedback (RLHF). However, overtly racist outputs have become rare, while more subtle, context-dependent racial stereotypes and assimilationist assumptions persist — appearing infrequently, often under specific linguistic or topical triggers. These rare stereotypical responses...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: e4a8565c-25eb-58c0-9b21-43ec145fd2c4
DOI: 10.14293/pr2199.003531.v2

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Finding Rare Racial Stereotypes in LLM Responses via Adversarial Anomaly DetectionDOI 10.14293/pr2199.003531.v2Read

Finding Rare Racial Stereotypes in LLM Responses via Adversarial Anomaly DetectionSemantic proximity Read

Discovering Latent Sycophantic Patterns in LLMs via Distributional Anomaly DetectionSemantic proximity Read

Discovering Latent Sycophantic Patterns in LLMs via Distributional Anomaly DetectionSemantic proximity Read

The Cognitive Fingerprint Problem: Transformer-LSTM Perplexity Geometry for Fair and Adversarially Robust AI Text DetectionSemantic proximity Read

Negative Preference Reduction in Large Language Model Unlearning: An Experimental ApproachSemantic proximity Read

On the Development of ToxicBias-Reasoning for Responsible Multicultural Bias Detection and ExplanationSemantic proximity Read

Bias and Cyberbullying Detection and Data Generation with Transformer AI Models and top LLMsSemantic proximity Read

Synthetic Data Generation for Bias Mitigation in AI: A Literature Review on Generative AI and Knowledge-Driven MethodsSemantic proximity Read

SAS Multimetric Tribunal: Experimental Validation of a Three-Zone Structural Classification System for Hallucination Detection in AI-Generated TextSemantic proximity Read

Quantifying and Mitigating Gender Bias in Legal Large Language Models:A Counterfactual Fairness FrameworkSemantic proximity Read

Emergent Numeric Bias in Large Language Models: An Empirical Study on the Anomalous Recurrence of the Number 27 Across Independent SessionsSemantic proximity Read

LLM-as-Critic: Contrastive and Adversarial Strategies for Authentic Text VerificationSemantic proximity Read