mlsanityData to 14 Oct 2026

Per-Entity Bias Mapping for AI Visibility: Why Brand Mentions Require Entity-Specific Calibration

arXiv 2606.21595 ↗First posted 19 Jun 2026cs.CL cs.IRcited by 0

Not on the map. Why?

Abstract (arXiv)

AI-mediated answer systems increasingly determine how brands and organizations are represented to users. Existing approaches reduce visibility to mention rate or citation frequency. This paper argues that aggregate metrics are insufficient because entities exhibit systematically different AI visibility error profiles. We introduce Per-Entity Bias Mapping (PEBM): a ten-dimensional framework distinguishing raw from verified mentions. Three failure modes are identified: (1) underrepresented entities suffer invisibility due to weak knowledge graph presence; (2) large entities suffer the Brand Hallucination Paradox -- model familiarity creates stronger surfaces for plausible but incorrect completions; (3) CEE entities face a structural infrastructure gap across knowledge graphs, NER, and entity linking. A fourth dimension, Parametric-Retrieval Lag Asymmetry, describes divergence between retrieval-augmented and parametric memory update cycles. A full-scale empirical study (n=100 Hungarian B2B entities, 1,400 probe runs, 2,062 sources) finds Tier 1 brands produce 52.69% fabricated citations versus 37.87% for Tier 3 entities (+14.82 pp; p=1.67e-11), supporting the Brand Hallucination Paradox. Regulatory-framed queries elevate fabrication to 56.77% versus 37.59% baseline (+19.2 pp). We identify rejection-induced confabulation escalation: agentic quality filters function as hallucination accelerators in compliance contexts. We introduce ghost cartography as a unifying mechanism: entities in sparse latent regions produce confident output interpolated from neighboring dense regions, yielding a two-dimensional confabulation space (fabricated presence vs. frozen representation).

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 21 of its 45 references that are in the map. First posted on arXiv 19 Jun 2026.

Builds on

21 of 45 references.

References · 21

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  2. 2023Towards Understanding Sycophancy in Language Modelscited by 1,040
  3. 2023Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedbackcited by 759
  4. 2023Survey on Factuality in Large Language Models: Knowledge, Retrieval and Domain-Specificitycited by 206
  5. 2024Large Legal Fictions: Profiling Legal Hallucinations in Large Language Modelscited by 150
  6. 2024Do I Know This Entity? Knowledge Awareness and Hallucinations in Language Modelscited by 115
  7. 2024Towards Lifelong Learning of Large Language Models: A Surveycited by 64
  8. 2025Large Language Models Hallucination: A Comprehensive Surveycited by 49
  9. 2025Parametric Retrieval Augmented Generationcited by 38
  10. 2022Geographic and Geopolitical Biases of Language Modelscited by 29
  11. 2024WildHallucinations: Evaluating Long-form Factuality in LLMs with Real-World Entity Queriescited by 26
  12. 2024Correctness is not Faithfulness in RAG Attributionscited by 18
9 more
  1. 2023Fairness in Language Models Beyond English: Gaps and Challengescited by 16
  2. 2024On the Limitations of Large Language Models (LLMs): False Attributioncited by 13
  3. 2021Analyzing Race and Country of Citizenship Bias in Wikidatacited by 6
  4. 2026Detecting and Correcting Reference Hallucinations in Commercial LLMs and Deep Research Agentscited by 6
  5. 2026CheckIfExist: Detecting Citation Hallucinations in the Era of AI-Generated Contentcited by 5
  6. 2024SLM Meets LLM: Balancing Latency, Interpretability and Consistency in Hallucination Detectioncited by 4
  7. 2026Location Not Found: Exposing Implicit Local and Global Biases in Multilingual LLMscited by 4
  8. 2025Social Biases in Knowledge Representations of Wikidata separates Global North from Global Southcited by 3
  9. 2026Don't Measure Once: Measuring Visibility in AI Search (GEO)cited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Prominence-Stratified Failure Modes in Retrieval-Augmented Commercial Recommendation: A 37,000-Run Audit
  2. 2026Faithfulness-QA: A Counterfactual Entity Substitution Dataset for Training Context-Faithful RAG Models
  3. 2025On Entity Identification in Language Models
  4. 2026Towards Faithful Industrial RAG: A Reinforced Co-adaptation Framework for Advertising QA
  5. 2026When RAG Fails to Equalize: Geo-bias in Factual Question Answering over Public Companies
  6. 2025Business Entity Entropy ↗
  7. 2026Does Bielik Know What It Doesn't Know? Activation Dispersion Separates Entity Familiarity from Factual Reliability Across Model Scale
  8. 2026EvidenT: Building Trustworthy Enterprise Assistants through Evidence Groundedness and Traceability