aiqbench analysis
This study provides a rigorous computational audit of ClinicalBERT, revealing that representational bias in medical language models is driven primarily by internal model amplification rather than direct inheritance from training data distributions. The findings highlight critical risks where demographic descriptors disproportionately skew probability predictions for specific patient groups, particularly regarding agency attribution and behavioral evaluations. For the AI community, this underscores the necessity of auditing clinical NLP systems to ensure equitable decision support before deployment in high-stakes healthcare environments.
A Computational Audit of Demographic Association Encoding in ClinicalBERT Language Predictions
arXiv:2606.14460v1 Announce Type: new Abstract: Transformer-based clinical language models are increasingly integrated into high-stakes clinical decision support pipelines, yet the computational mechanisms through which demographic associations encoded in medical documentation propagate into model probability distributions remain empirically underspecified. We present a systematic computational audit of representational bias in ClinicalBERT (Alsentzer et al., 2019), a BERT-based model pretrained on MIMIC-III discharge summaries, employing two complementary probing methodologies: Log Probability Bias Analysis (LPBA), which quantifies demographic descriptor-induced shifts in masked token probability distributions across behavioral and evaluative semantic categories, and Masked Language Model-based analysis (MLM), which probes internal representational structure for demographic agency attribution encoding across 98 real clinical sentence templates and eight intersectional race-gender combinations. Corpus frequency analysis operationalizes the distinction between statistical disparity and bias amplification by benchmarking model outputs against empirical term frequencies in the MIMIC-III training corpus. Of 32 statistically significant findings, 65.6% contradict observed corpus distributions, rising to 80% for Black patients and 87.5% for agency attribution under MLM probing, providing direct empirical evidence that representational bias in ClinicalBERT operates predominantly through model-internal amplification rather than training data inheritance. Keywords: natural language processing, clinical documentation, algorithmic auditing, representational bias, health equity 1
Abstract:Transformer-based clinical language models are increasingly integrated into high-stakes clinical decision support pipelines, yet the computational mechanisms through which demographic associations encoded in medical documentation propagate into model probability distributions remain empirically underspecified. We present a systematic computational audit of representational bias in ClinicalBERT (Alsentzer et al., 2019), a BERT-based model pretrained on MIMIC-III discharge summaries, employing two complementary probing methodologies: Log Probability Bias Analysis (LPBA), which quantifies demographic descriptor-induced shifts in masked token probability distributions across behavioral and evaluative semantic categories, and Masked Language Model-based analysis (MLM), which probes internal representational structure for demographic agency attribution encoding across 98 real clinical sentence templates and eight intersectional race-gender combinations. Corpus frequency analysis operationalizes the distinction between statistical disparity and bias amplification by benchmarking model outputs against empirical term frequencies in the MIMIC-III training corpus. Of 32 statistically significant findings, 65.6% contradict observed corpus distributions, rising to 80% for Black patients and 87.5% for agency attribution under MLM probing, providing direct empirical evidence that representational bias in ClinicalBERT operates predominantly through model-internal amplification rather than training data inheritance. Keywords: natural language processing, clinical documentation, algorithmic auditing, representational bias, health equity 1
Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
arXiv:2606.14460 [cs.CL]
(or arXiv:2606.14460v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2606.14460 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 288 words
Related coverage
Fairness Beyond Anonymization? Demographic Leakage in German LLM-Generated Resumes
arXiv · Sep 22, 2026
AWARE-FX: An Auditable Knowledge-Guided AI System for Measuring Corporate Foreign-Exchange Hedging Disclosure
arXiv · Jul 31, 2026
J++ Lens: Jacobian Filtering Enables More Faithful Workspace Lenses
LessWrong · Oct 8, 2026
Always ask your agent to be honest
LessWrong · Sep 16, 2026