Inspired by recent plug-in frameworks that repurpose frozen layers from large language models (LLMs) as inductive priors, we explore whether such mechanisms can be extended to clinical time-series prediction without textual inputs or LLM fine-tuning. We introduce a lightweight plug-in architecture that inserts a single frozen LLM Transformer layer between an aggregated time-series representation and the prediction head. Unlike prior work focused on vision or language tasks, our study targets clinical time-series data, where LLMs typically underperform when applied directly.Experiments on two ICU prediction tasks from MIMIC-III show that the proposed plug-in exhibits heterogeneous effects across different backbones and tasks, with occasional performance improvements and minimal computational overhead. We further compare general-purpose and medical-domain LLM layers under an identical plug-in setting, analyzing how domain specialization interacts with clinical time-series models. Overall, our results highlight important limitations of frozen LLM plug-ins and motivate future work on understanding the conditions under which such layers may be beneficial.

Electronic medical record (EMR) databases are widely used for clinical machine learning, but their preprocessing remains difficult to reproduce because cohort definitions, patient-level features, and clinical item-code mappings must be adapted to each database schema. We formulate EMR preprocessing as the task of translating study requests into database-specific cohort/feature extraction SQL and local clinical item-code mappings. Existing preprocessing and harmonization frameworks remain difficult to adapt because they rely on dataset-specific rules or predefined input formats. We introduce EMR-AGENT, an LLM-agent framework that reduces reliance on manual rule writing by automating database-specific cohort selection, feature extraction, and clinical code mapping from study requests. EMR-AGENT retrieves schema and documentation context, interacts with live databases through SQL, observes execution results, and refines extraction decisions from error feedback. To enable rigorous evaluation, we develop a benchmarking codebase for three intensive care unit (ICU) databases (MIMIC-III, eICU, and SICdb), covering reference databases and a newer SICdb version released after the backbone’s knowledge cutoff. Across these benchmarks, EMR-AGENT reduces schema-debugging and item-code search effort, enabling reproducible evaluation of preprocessing quality under expert review, while reducing reliance on hand-written database-specific pipelines.

Peer review serves as a backbone of academic research, but in most AI conferences, the review quality is degrading as the number of submissions explodes. To reliably detect low-quality reviews, we define misinformed review points as either "weaknesses" in a review that contain incorrect premises, or "questions" in a review that can be already answered by the paper. We verify that 15.2% of weaknesses and 26.4% of questions are misinformed and introduce ReviewScore indicating if a review point is misinformed. To evaluate the factuality of each premise of weaknesses, we propose an automated engine that reconstructs every explicit and implicit premise from a weakness. We build a human expert-annotated ReviewScore dataset to check the ability of LLMs to automate ReviewScore evaluation. Then, we measure human-model agreements on ReviewScore using eight current state-of-the-art LLMs. The models show F1 scores of 0.4–0.5 and kappa scores of 0.3–0.4, indicating moderate agreement but also suggesting that fully automating the evaluation remains challenging. A thorough disagreement analysis reveals that most errors are due to models’ incorrect reasoning. We also prove that evaluating premise-level factuality shows significantly higher agreements than evaluating weakness-level factuality.