The paper identifies a new problem in clinical natural language processing called the clinical lost‑in‑the‑middle (CLitM) effect, where large language models perform poorly on information located near the center of long electronic health record (EHR) documents. Using the MedAlign dataset, the authors quantify a 21.9‑percentage‑point accuracy gap across 2,196 instruction‑response pairs and six models, showing that most critical facts lie in the CLitM trough. They propose Query‑Conditioned Clinical Suppression (QCCS), a lightweight context‑selection gate that outperforms traditional retrieval methods (BM25, dense retrieval, cross‑encoder reranking) on a held‑out set of 83 instructions, achieving up to 25.3% accuracy for middle‑position queries. whyItMatters":"The study demonstrates that standard retrieval strategies fail to reliably surface central clinical information, and that a query‑aligned selection mechanism can substantially improve model performance on critical EHR data."
arXiv:2609.14825v1 Announce Type: cross Abstract: Large language models (LLMs) are often deemed unsafe for clinical question answering because of their tendency to hallucinate. Retrieval augmentation...
arXiv:2610.34871v1 Announce Type: new Abstract: Objective: The rapid advancement of large language models (LLMs) in healthcare has opened new possibilities for AI-assisted clinical decision-making; h...
arXiv:2508. 14817v2 Announce Type: replace-cross Abstract: Objective: To evaluate whether retrieval-augmented generation (RAG) can serve as an efficient alternative to long-context prompting for clinical reasoning over electronic health records (EHRs).
The paper proposes a training‑free detector that uses sentence‑level context sensitivity to identify unsupported content in retrieval‑augmented generation (RAG) answers. By re‑scoring each sentence with full context, no context, and each chunk removed, the method flags the chunk whose removal most reduces a sentence’s likelihood as the likely source. Evaluated on RAGTruth, TofuEval, and RAGBench, the detector outperforms answer‑level faithfulness scores, achieving AUCs up to 0.745 and matching per‑chunk fact‑checkers while using only a fraction of the compute required by large‑language‑model judges.
arXiv:2606. 16890v1 Announce Type: cross Abstract: Aggregate accuracy benchmarks conceal a systematic structure in how large language models fail at electronic health record (EHR) question answering: questions requiring more inferential steps produce disproportionately more errors.
The paper evaluates whether a 9‑billion parameter decision foundation model, Clef‑Flash, is necessary for symptom‑to‑disease classification. Fine‑tuning the model on the 24‑class Symptom2Disease benchmark yields 95.42% accuracy, but simple TF‑IDF linear classifiers achieve higher accuracy (98.75%) with far faster inference. Ablation studies show that predictions can hinge on single descriptive clauses, indicating that the benchmark may not fully capture the model’s capabilities.
The paper introduces Clinical Intent Extraction (CIE), a task that transforms fragmented clinical action annotations into complete structured records called Clinical Intent Representation (CIR). CIR decomposes each action into verb, type, coded target, timing, condition, request‑intent (aligned to HL7 FHIR) and modality, adding dimensions absent in prior datasets. By re‑expressing five heterogeneous corpora into CIR, the authors create CIRCA, a benchmark of 10,011 harmonized intents with human‑validated subsets, crosswalks, and a deterministic FHIR R4 mapper, and demonstrate that existing models perform poorly on the full task, highlighting the need for targeted development.
arXiv:2607. 04223v1 Announce Type: cross Abstract: Retrieval-augmented generation (RAG) reduces but does not eliminate hallucination, and existing detectors return a single answer-level score that does not indicate which sentence is unsupported, or why.
arXiv:2511.00421v2 Announce Type: replace-cross Abstract: Large language models (LLMs) show promise in medical applications, but their ability to detect and correct errors in clinical texts remains u...
TCM-DiagCommit is a new benchmark derived from the TCM-SD/ZY-BERT dataset that expands 1,834 source cases into 23,842 task rows covering 148 syndrome labels. It evaluates models on three diagnostic behaviors—distinguishing semantic-neighbor syndromes, updating predictions as staged evidence (t1/t2/t3) is revealed, and committing under uncertainty—using a benchmark-defined action contract that scores commit, acquire, defer, and abstain. Experiments with five zero‑shot DirectLLM calibrations show that while random10 top‑1 accuracy ranges from 63.2% to 66.4%, performance drops sharply on harder subsets (hard20 and full148), highlighting that final‑answer accuracy alone is insufficient for diagnostic evaluation.
The paper introduces a retrieval‑augmented multi‑agent framework that automatically generates instance‑specific evaluation rubrics for medical language models. By retrieving authoritative medical evidence, decomposing it into atomic facts, and combining these with user interaction constraints, the system produces fine‑grained criteria that outperform GPT‑4o on HealthBench and LLMEval‑Med. The generated rubrics also guide response refinement, improving medical LLM output quality by 9.2%.
arXiv:2604. 04593v2 Announce Type: replace-cross Abstract: Retrieval-augmented generation (RAG) grounds large language models in external medical knowledge, yet standard retrievers frequently surface hard negatives that are semantically close to the query but describe clinically distinct conditions.