Metraipolis › archive › dc_1f48e27fdb13

Inhibitory Attention for Clinical Long-Context Reasoning: Characterizing and Mitigating Lost-in-the-Middle Effects in EHR Processing — arXiv:2608.20348

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.20348
License (as archived)
CC0-1.0
Archived
2026-10-18T10:18:05.347Z
SHA-256 of the archived text
df29f9325eed40390e2f59dbcded9a166b50cb90ed131a7652ee996946376251
Stable URL of this page
https://metraipolis.com/documento/dc_1f48e27fdb13

How to cite

Inhibitory Attention for Clinical Long-Context Reasoning: Characterizing and Mitigating Lost-in-the-Middle Effects in EHR Processing — arXiv:2608.20348 — arXiv (en), https://arxiv.org/abs/2608.20348 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1f48e27fdb13 (Metraipolis archive dc_1f48e27fdb13, sha256 df29f9325eed40390e2f59dbcded9a166b50cb90ed131a7652ee996946376251)
BibTeX
@misc{dc1f48e27fdb13,
  title = {Inhibitory Attention for Clinical Long-Context Reasoning: Characterizing and Mitigating Lost-in-the-Middle Effects in EHR Processing — arXiv:2608.20348},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.20348},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1f48e27fdb13; sha256 df29f9325eed40390e2f59dbcded9a166b50cb90ed131a7652ee996946376251},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.20348. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Electronic health records now routinely exceed 100,000 tokens per patient. Yet large language models exhibit the lost-in-the-middle (LitM) effect: information near the center of a long context is retrieved less reliably than information near the edges. In clinical use this is not benign: the single most consequential fact in a note can sit at its center. We term this the clinical lost-in-the-middle (CLitM) problem, give its first systematic characterization using MedAlign, and compare context-selection strategies as remedies. Across 2,196 instruction-response pairs and six language models, we observe a 21.9 percentage-point gap between peak accuracy (59.5%, 95% CI [46.3, 71.0], 20-30% decile) and trough accuracy (37.6% [23.2, 52.5] at 70-80%); 67.8% of reference answers fall between the 10th and 90th percentiles of the EHR timeline, inside the CLitM trough. We introduce Query-Conditioned Clinical Suppression (QCCS), a lightweight query-conditioned selection gate, and evaluate it against BM25, BM25 with section-header filtering, dense retrieval, and cross-encoder reranking (N=83 held-out instructions). With Qwen2.5-7B-Instruct (16k context), QCCS outperforms all five comparators under LLM-as-judge scoring: for middle-position instructions QCCS reaches 16.7% versus BM25 3.3%, cross-encoder 0.0%, dense 0.0%, and full context 6.7%; overall QCCS reaches 25.3% versus at most 3.6% for retrieval-only comparators. This advantage is not explained by retrieval recall: at k=20, BM25 retrieves the gold evidence sentence in 98.8% of instructions (QCCS 34.9%), yet retrieval arms stay at most 2.6% accurate even when they retrieve it, whereas QCCS reaches 25.0% even when it does not. In this proof-of-concept evaluation, query-aligned context selection predicts EHR instruction-following accuracy better than gold-sentence retrieval recall.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-24 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.20348 — 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 2026-10-18T10:17:37.544Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-24?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).