FairLLaVA: Fairness-Aware Parameter-Efficient Fine-Tuning for Large Vision-Language Assistants

While powerful in image-conditioned generation, multimodal large language models (MLLMs) can display uneven performance across demographic groups, highlighting fairness risks. In safety-critical clinical settings, such disparities risk producing unequal diagnostic narratives and eroding trust in AI-assisted decision-making. While fairness has been studied extensively in vision-only and language-only models, its impact on MLLMs remains largely underexplored. To address these biases, we introduce FairLLaVA, a parameter-efficient fine-tuning method that mitigates group disparities in visual instruction tuning without compromising overall performance. By minimizing the mutual information between target attributes, FairLLaVA regularizes the model's representations to be demographic-invariant. The method can be incorporated as a lightweight plug-in, maintaining efficiency with low-rank adapter fine-tuning, and provides an architecture-agnostic approach to fair visual instruction following. Extensive experiments on large-scale chest radiology report generation and dermoscopy visual question answering benchmarks show that FairLLaVA consistently reduces inter-group disparities while improving both equity-scaled clinical performance and natural language generation quality across diverse medical imaging modalities. Code can be accessed at https://github.com/bhosalems/FairLLaVA.

Automating Clinical Information Retrieval from Finnish Electronic Health Records Using Large Language Models

Clinicians often need to retrieve patient-specific information from electronic health records (EHRs), a task that is time-consuming and error-prone. We present a locally deployable Clinical Contextual Question Answering (CCQA) framework that answers clinical questions directly from EHRs without external data transfer. Open-source large language models (LLMs) ranging from 4B to 70B parameters were benchmarked under fully offline conditions using 1,664 expert-annotated question-answer pairs derived from records of 183 patients. The dataset consisted predominantly of Finnish clinical text. In free-text generation, Llama-3.1-70B achieved 95.3% accuracy and 97.3% consistency across semantically equivalent question variants, while the smaller Qwen3-30B-A3B-2507 model achieved comparable performance. In a multiple-choice setting, models showed similar accuracy but variable calibration. Low-precision quantization (4-bit and 8-bit) preserved predictive performance while reducing GPU memory requirements and improving deployment feasibility. Clinical evaluation identified clinically significant errors in 2.9% of outputs, and semantically equivalent questions occasionally yielded discordant responses, including instances where one formulation was correct and the other contained a clinically significant error (0.96% of cases). These findings demonstrate that locally hosted open-source LLMs can accurately retrieve patient-specific information from EHRs using natural-language queries, while highlighting the need for validation and human oversight in clinical deployment.

Why Models Know But Don't Say: Chain-of-Thought Faithfulness Divergence Between Thinking Tokens and Answers in Open-Weight Reasoning Models

Extended-thinking models expose a second text-generation channel ("thinking tokens") alongside the user-visible answer. This study examines 12 open-weight reasoning models on MMLU and GPQA questions paired with misleading hints. Among the 10,506 cases where models actually followed the hint (choosing the hint's target over the ground truth), each case is classified by whether the model acknowledges the hint in its thinking tokens, its answer text, both, or neither. In 55.4% of these cases the model's thinking tokens contain hint-related keywords that the visible answer omits entirely, a pattern termed *thinking-answer divergence*. The reverse (answer-only acknowledgment) is near-zero (0.5%), confirming that the asymmetry is directional. Hint type shapes the pattern sharply: sycophancy is the most *transparent* hint, with 58.8% of sycophancy-influenced cases acknowledging the professor's authority in both channels, while consistency (72.2%) and unethical (62.7%) hints are dominated by thinking-only acknowledgment. Models also vary widely, from near-total divergence (Step-3.5-Flash: 94.7%) to relative transparency (Qwen3.5-27B: 19.6%). These results show that answer-text-only monitoring misses more than half of all hint-influenced reasoning and that thinking-token access, while necessary, still leaves 11.8% of cases with no verbalized acknowledgment in either channel.

Beyond Code Snippets: Benchmarking LLMs on Repository-Level Question Answering

Large Language Models (LLMs) have shown impressive capabilities across software engineering tasks, including question answering (QA). However, most studies and benchmarks focus on isolated functions or single-file snippets, overlooking the challenges of real-world program comprehension, which often spans multiple files and system-level dependencies. In this work, we introduce StackRepoQA, the first multi-project, repository-level question answering dataset constructed from 1,318 real developer questions and accepted answers across 134 open-source Java projects. Using this dataset, we systematically evaluate two widely used LLMs (Claude 3.5 Sonnet and GPT-4o) under both direct prompting and agentic configurations. We compare baseline performance with retrieval-augmented generation methods that leverage file-level retrieval and graph-based representations of structural dependencies. Our results show that LLMs achieve moderate accuracy at baseline, with performance improving when structural signals are incorporated. Nonetheless, overall accuracy remains limited for repository-scale comprehension. The analysis reveals that high scores often result from verbatim reproduction of Stack Overflow answers rather than genuine reasoning. To our knowledge, this is the first empirical study to provide such evidence in repository-level QA. We release StackRepoQA to encourage further research into benchmarks, evaluation protocols, and augmentation strategies that disentangle memorization from reasoning, advancing LLMs as reliable tool for repository-scale program comprehension.

AD-CARE: A Guideline-grounded, Modality-agnostic LLM Agent for Real-world Alzheimer's Disease Diagnosis with Multi-cohort Assessment, Fairness Analysis, and Reader Study

Alzheimer's disease (AD) is a growing global health challenge as populations age, and timely, accurate diagnosis is essential to reduce individual and societal burden. However, real-world AD assessment is hampered by incomplete, heterogeneous multimodal data and variability across sites and patient demographics. Although large language models (LLMs) have shown promise in biomedicine, their use in AD has largely been confined to answering narrow, disease-specific questions rather than generating comprehensive diagnostic reports that support clinical decision-making. Here we expand LLM capabilities for clinical decision support by introducing AD-CARE, a modality-agnostic agent that performs guideline-grounded diagnostic assessment from incomplete, heterogeneous inputs without imputing missing modalities. By dynamically orchestrating specialized diagnostic tools and embedding clinical guidelines into LLM-driven reasoning, AD-CARE generates transparent, report-style outputs aligned with real-world clinical workflows. Across six cohorts comprising 10,303 cases, AD-CARE achieved 84.9% diagnostic accuracy, delivering 4.2%-13.7% relative improvements over baseline methods. Despite cohort-level differences, dataset-specific accuracies remain robust (80.4%-98.8%), and the agent consistently outperforms all baselines. AD-CARE reduced performance disparities across racial and age subgroups, decreasing the average dispersion of four metrics by 21%-68% and 28%-51%, respectively. In a controlled reader study, the agent improved neurologist and radiologist accuracy by 6%-11% and more than halved decision time. The framework yielded 2.29%-10.66% absolute gains over eight backbone LLMs and converges their performance. These results show that AD-CARE is a scalable, practically deployable framework that can be integrated into routine clinical workflows for multimodal decision support in AD.

Dialogue to Question Generation for Evidence-based Medical Guideline Agent Development

Evidence-based medicine (EBM) is central to high-quality care, but remains difficult to implement in fast-paced primary care settings. Physicians face short consultations, increasing patient loads, and lengthy guideline documents that are impractical to consult in real time. To address this gap, we investigate the feasibility of using large language models (LLMs) as ambient assistants that surface targeted, evidence-based questions during physician-patient encounters. Our study focuses on question generation rather than question answering, with the aim of scaffolding physician reasoning and integrating guideline-based practice into brief consultations. We implemented two prompting strategies, a zero-shot baseline and a multi-stage reasoning variant, using Gemini 2.5 as the backbone model. We evaluated on a benchmark of 80 de-identified transcripts from real clinical encounters, with six experienced physicians contributing over 90 hours of structured review. Results indicate that while general-purpose LLMs are not yet fully reliable, they can produce clinically meaningful and guideline-relevant questions, suggesting significant potential to reduce cognitive burden and make EBM more actionable at the point of care.

Reaching Beyond the Mode: RL for Distributional Reasoning in Language Models

Given a question, a language model (LM) implicitly encodes a distribution over possible answers. In practice, post-training procedures for LMs often collapse this distribution onto a single dominant mode. While this is generally not a problem for benchmark-style evaluations that assume one correct answer, many real-world tasks inherently involve multiple valid answers or irreducible uncertainty. Examples include medical diagnosis, ambiguous question answering, and settings with incomplete information. In these cases, we would like LMs to generate multiple plausible hypotheses, ideally with confidence estimates for each one, and without computationally intensive repeated sampling to generate non-modal answers. This paper describes a multi-answer reinforcement learning approach for training LMs to perform distributional reasoning over multiple answers during inference. We modify the RL objective to enable models to explicitly generate multiple candidate answers in a single forward pass, internalizing aspects of inference-time search into the model's generative process. Across question-answering, medical diagnostic, and coding benchmarks, we observe improved diversity, coverage, and set-level calibration scores compared to single answer trained baselines. Models trained with our approach require fewer tokens to generate multiple answers than competing approaches. On coding tasks, they are also substantially more accurate. These results position multi-answer RL as a principled and compute-efficient alternative to inference-time scaling procedures such as best-of-k. Code and more information can be found at https://multi-answer-rl.github.io/.

Multi-Agent Reasoning with Consistency Verification Improves Uncertainty Calibration in Medical MCQA

Miscalibrated confidence scores are a practical obstacle to deploying AI in clinical settings. A model that is always overconfident offers no useful signal for deferral. We present a multi-agent framework that combines domain-specific specialist agents with Two-Phase Verification and S-Score Weighted Fusion to improve both calibration and discrimination in medical multiple-choice question answering. Four specialist agents (respiratory, cardiology, neurology, gastroenterology) generate independent diagnoses using Qwen2.5-7B-Instruct. Each diagnosis is then subjected to a two-phase self-verification process that measures internal consistency and produces a Specialist Confidence Score (S-score). The S-scores drive a weighted fusion strategy that selects the final answer and calibrates the reported confidence. We evaluate across four experimental settings, covering 100-question and 250-question high-disagreement subsets of both MedQA-USMLE and MedMCQA. Calibration improvement is the central finding, with ECE reduced by 49-74% across all four settings, including the harder MedMCQA benchmark where these gains persist even when absolute accuracy is constrained by knowledge-intensive recall demands. On MedQA-250, the full system achieves ECE = 0.091 (74.4% reduction over the single-specialist baseline) and AUROC = 0.630 (+0.056) at 59.2% accuracy. Ablation analysis identifies Two-Phase Verification as the primary calibration driver and multi-agent reasoning as the primary accuracy driver. These results establish that consistency-based verification produces more reliable uncertainty estimates across diverse medical question types, providing a practical confidence signal for deferral in safety-critical clinical AI applications.

MedObvious: Exposing the Medical Moravec's Paradox in VLMs via Clinical Triage

Vision Language Models (VLMs) are increasingly used for tasks like medical report generation and visual question answering. However, fluent diagnostic text does not guarantee safe visual understanding. In clinical practice, interpretation begins with pre-diagnostic sanity checks: verifying that the input is valid to read (correct modality and anatomy, plausible viewpoint and orientation, and no obvious integrity violations). Existing benchmarks largely assume this step is solved, and therefore miss a critical failure mode: a model can produce plausible narratives even when the input is inconsistent or invalid. We introduce MedObvious, a 1,880-task benchmark that isolates input validation as a set-level consistency capability over small multi-panel image sets: the model must identify whether any panel violates expected coherence. MedObvious spans five progressive tiers, from basic orientation/modality mismatches to clinically motivated anatomy/viewpoint verification and triage-style cues, and includes five evaluation formats to test robustness across interfaces. Evaluating 17 different VLMs, we find that sanity checking remains unreliable: several models hallucinate anomalies on normal (negative-control) inputs, performance degrades when scaling to larger image sets, and measured accuracy varies substantially between multiple-choice and open-ended settings. These results show that pre-diagnostic verification remains unsolved for medical VLMs and should be treated as a distinct, safety-critical capability before deployment.

Mixture of Demonstrations for Textual Graph Understanding and Question Answering

Textual graph-based retrieval-augmented generation (GraphRAG) has emerged as a powerful paradigm for enhancing large language models (LLMs) in domain-specific question answering. While existing approaches primarily focus on zero-shot GraphRAG, selecting high-quality demonstrations is crucial for improving reasoning and answer accuracy. Furthermore, recent studies have shown that retrieved subgraphs often contain irrelevant information, which can degrade reasoning performance. In this paper, we propose MixDemo, a novel GraphRAG framework enhanced with a Mixture-of-Experts (MoE) mechanism for selecting the most informative demonstrations under diverse question contexts. To further reduce noise in the retrieved subgraphs, we introduce a query-specific graph encoder that selectively attends to information most relevant to the query. Extensive experiments across multiple textual graph benchmarks show that MixDemo significantly outperforms existing methods.