Metraipolis › archive › dc_dfe52c22b375

Auditing evaluation artifacts and data leakage in large language model systems for rare disease diagnosis — arXiv:2610.33592

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.33592
License (as archived): CC0-1.0
Archived: 2026-10-27T07:56:39.914Z
SHA-256 of the archived text: 42e6ff1ba1ec6268ea74eed6ee9363aaa6b146def3085cc3298c39094c9a7894
Stable URL of this page: https://metraipolis.com/documento/dc_dfe52c22b375

How to cite

Auditing evaluation artifacts and data leakage in large language model systems for rare disease diagnosis — arXiv:2610.33592 — arXiv (en), https://arxiv.org/abs/2610.33592 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_dfe52c22b375 (Metraipolis archive dc_dfe52c22b375, sha256 42e6ff1ba1ec6268ea74eed6ee9363aaa6b146def3085cc3298c39094c9a7894)

BibTeX

@misc{dcdfe52c22b375,
  title = {Auditing evaluation artifacts and data leakage in large language model systems for rare disease diagnosis — arXiv:2610.33592},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.33592},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_dfe52c22b375; sha256 42e6ff1ba1ec6268ea74eed6ee9363aaa6b146def3085cc3298c39094c9a7894},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.33592. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Augmented large language model (LLM) systems have reported high benchmark accuracies in rare disease diagnosis, yet their real-world clinical utility remains unproven. A recent meta-analysis flagged augmented LLM diagnostic evaluations as having a high risk of bias due to suspected data leakage; however, the precise impact of these vulnerabilities on model inference has yet to be fully quantified. In this study, we present a comprehensive empirical audit to quantify evaluation artifacts across prevailing AI paradigms, using a repository of over 60,000 cases from diverse public and independent clinical cohorts. By reconstructing a state-of-the-art agentic framework, we demonstrate that candidate-space contamination artificially inflates performance: enforcing strict cross-dataset isolation resulted in a marked performance decline, with the cross-dataset mean Recall@1 decreasing from 54.3% (under conventional one-vs-rest evaluation) to 40.3% (Hedges' g = 0.94). Similar spatial overfitting in supervised classifiers caused diagnostic accuracy to drop from an internal mean of 42.9% to 16.7% across external cohorts (Cohen's d = 2.01). Furthermore, we quantify how LLMs infer diagnoses by exploiting retrospective therapeutic cues in clinical narratives, and how temporal misalignment in foundational knowledge bases leads to performance fluctuations. We conclude that prevailing evaluation protocols for augmented LLM systems inadvertently reflect data leakage rather than genuine medical reasoning. To address these systematic artifacts, we propose a rigorous evaluation framework - coupling strict data-isolation protocols with standardized benchmarking safeguards - to ensure future diagnostic AI for rare diseases is evaluated on actual clinical robustness.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.33592 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-27T07:55:30.646Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).