Preprint · 2026 · Open access

CLExEval: A Human-in-the-Loop Framework for Qualitative Evaluation of LLM Clinical Reasoning

Ajmal M., Abin Roy, Afthab Salam Kanniyan, Jawadh Abdul Kabeer, Jerin James, Preslav Nakov, Zhuohan Xie

arXiv (Cornell University) · Published 30 June 2026 · DOI 10.48550/arxiv.2606.31608

Open PDFView at the publisherCite this paper

In brief

“CLExEval: A Human-in-the-Loop Framework for Qualitative Evaluation of LLM Clinical Reasoning” is a 2026 preprint by Ajmal M. and 6 co-authors, published in arXiv (Cornell University). OpenAlex counts 0 citations as of 19 November 2026. It is open access under a CC BY licence: the full text is free to read at arXiv (Cornell University).

Citations
0
References
36
Authors
7

Abstract

Large Language Models (LLMs) achieve strong results on many medical benchmarks, but their clinical reasoning remains difficult to evaluate reliably. A central risk is an evaluation illusion: fluent and well-structured explanations can appear clinically convincing even when the final diagnosis is incorrect. We introduce CLExEval, a human-in-the-loop framework for evaluating LLM clinical reasoning under progressive information masking. CLExEval combines 5,600 expert-physician annotations with 200 clinical reasoning traces derived from 40 rare diagnostic cases. Our analysis identifies three recurring failure patterns: (i) verbosity bias, where GPT-4o-mini's diagnostic accuracy drops from 95.0% to 32.5% under information scarcity; (ii) a hidden knowledge paradox, where a specialist model reaches 92.5% maximum diagnostic potential but fails to retrieve that knowledge reliably in verbose contexts; and (iii) a 68.6% reasoning-to-output mismatch, where correct diagnoses appear in reasoning traces but are not reflected in final answers. We further evaluate the LLM-as-a-Judge paradigm on a human-verified failure set (n = 142). GPT-4o-mini approved 47.9% of clinically incorrect outputs, while HuatuoGPT-o1 approved all validly scored failures and showed a positive self-preference bias. These results suggest that standalone automated clinical evaluations can substantially overestimate clinical reliability without expert-grounded validation.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

M., A., Roy, A., Kanniyan, A. S., Kabeer, J. A., James, J., Nakov, P., & Xie, Z. (2026). CLExEval: A Human-in-the-Loop Framework for Qualitative Evaluation of LLM Clinical Reasoning. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2606.31608

MLA 9

M., Ajmal, et al. “CLExEval: A Human-in-the-Loop Framework for Qualitative Evaluation of LLM Clinical Reasoning.” arXiv (Cornell University), 2026. https://doi.org/10.48550/arxiv.2606.31608.

Chicago

M., Ajmal, Abin Roy, Afthab Salam Kanniyan, Jawadh Abdul Kabeer, Jerin James, Preslav Nakov, and Zhuohan Xie. 2026. “CLExEval: A Human-in-the-Loop Framework for Qualitative Evaluation of LLM Clinical Reasoning.” arXiv (Cornell University). https://doi.org/10.48550/arxiv.2606.31608.

Harvard

M., A. et al. (2026) ‘CLExEval: A Human-in-the-Loop Framework for Qualitative Evaluation of LLM Clinical Reasoning’, arXiv (Cornell University). doi:10.48550/arxiv.2606.31608.

IEEE

A. M. et al., “CLExEval: A Human-in-the-Loop Framework for Qualitative Evaluation of LLM Clinical Reasoning,” arXiv (Cornell University), 2026, doi: 10.48550/arxiv.2606.31608.

BibTeX

@misc{m2026clexeval,
  title     = {{CLExEval: A Human-in-the-Loop Framework for Qualitative Evaluation of LLM Clinical Reasoning}},
  author    = {M., Ajmal and Roy, Abin and Kanniyan, Afthab Salam and Kabeer, Jawadh Abdul and James, Jerin and Nakov, Preslav and Xie, Zhuohan},
  howpublished = {arXiv (Cornell University)},
  year      = {2026},
  doi       = {10.48550/arxiv.2606.31608},
  url       = {https://doi.org/10.48550/arxiv.2606.31608}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Topic and keywords

Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education

  • large language models
  • clinical reasoning
  • human-in-the-loop evaluation
  • information masking
  • diagnostic accuracy
  • LLM-as-a-judge
  • GPT-4o-mini

Work with this paper

Most-cited papers in Medicine

All highly cited Medicine papers

Questions about this paper

How do I cite this paper in APA style?

M., A., Roy, A., Kanniyan, A. S., Kabeer, J. A., James, J., Nakov, P., & Xie, Z. (2026). CLExEval: A Human-in-the-Loop Framework for Qualitative Evaluation of LLM Clinical Reasoning. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2606.31608

How many times has this paper been cited?

0 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.

Is this paper free to read?

Yes. It is open access at arXiv (Cornell University): sign in and open the PDF from this page.

Who wrote this paper?

Ajmal M., Abin Roy, Afthab Salam Kanniyan, Jawadh Abdul Kabeer, Jerin James, Preslav Nakov, Zhuohan Xie.

What is the DOI of this paper?

10.48550/arxiv.2606.31608, permanently at https://doi.org/10.48550/arxiv.2606.31608. Sign in and open it from this page.

Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.