Med-StepBench: A Hierarchical Reasoning Framework for Evaluating Hallucinations in Medical Vision-Language Models

Large vision-language models (VLMs) demonstrate strong performance in medical image understanding, but frequently generate clinically plausible yet incorrect statements, raising significant safety concerns. Existing medical hallucination benchmarks primarily focus on 2D imaging with one-shot diagnostic questions, offering limited insight into whether predictions are grounded in correct localization and abnormality identification, allowing critical reasoning errors to remain hidden behind seemingly correct diagnoses. We introduce Med-StepBench, the first large-scale benchmark for step-wise hallucination detection in 3D oncological PET/CT, comprising over 12,000 images and more than 1,000,000 image-statement pairs across volumetric and multi-view 2D data, which decomposes clinical reasoning into four expert-designed diagnostic stages. Using clinician-verified annotations, we perform the first step-level evaluation of general-purpose and medical VLMs, revealing systematic failure modes obscured by aggregate accuracy metrics. Furthermore, we show that current VLMs are highly susceptible to adversarial yet clinically plausible intermediate explanations, which significantly amplify hallucinations despite contradictory visual evidence. Together, our findings highlight fundamental limitations in grounding multi-step clinical reasoning and establish Med-StepBench as a rigorous benchmark for developing safer and more reliable medical VLMs.

Region-Grounded Report Generation for 3D Medical Imaging: A Fine-Grained Dataset and Graph-Enhanced Framework

Automated medical report generation for 3D PET/CT imaging is fundamentally challenged by the high-dimensional nature of volumetric data and a critical scarcity of annotated datasets, particularly for low-resource languages. Current black-box methods map whole volumes to reports, ignoring the clinical workflow of analyzing localized Regions of Interest (RoIs) to derive diagnostic conclusions. In this paper, we bridge this gap by introducing VietPET-RoI, the first large-scale 3D PET/CT dataset with fine-grained RoI annotation for a low-resource language, comprising 600 PET/CT samples and 1,960 manually annotated RoIs, paired with corresponding clinical reports. Furthermore, to demonstrate the utility of this dataset, we propose HiRRA, a novel framework that mimics the professional radiologist diagnostic workflow by employing graph-based relational modules to capture dependencies between RoI attributes. This approach shifts from global pattern matching toward localized clinical findings. Additionally, we introduce new clinical evaluation metrics, namely RoI Coverage and RoI Quality Index, that measure both RoI localization accuracy and attribute description fidelity using LLM-based extraction. Extensive evaluation demonstrates that our framework achieves SOTA performance, surpassing existing models by 19.7% in BLEU and 4.7% in ROUGE-L, while achieving a remarkable 45.8% improvement in clinical metrics, indicating enhanced clinical reliability and reduced hallucination. Our code and dataset are available on GitHub.

Beyond the Global Scores: Fine-Grained Token Grounding as a Robust Detector of LVLM Hallucinations

Large vision-language models (LVLMs) achieve strong performance on visual reasoning tasks but remain highly susceptible to hallucination. Existing detection methods predominantly rely on coarse, whole-image measures of how an object token relates to the input image. This global strategy is limited: hallucinated tokens may exhibit weak but widely scattered correlations across many local regions, which aggregate into deceptively high overall relevance, thus evading the current global hallucination detectors. We begin with a simple yet critical observation: a faithful object token must be strongly grounded in a specific image region. Building on this insight, we introduce a patch-level hallucination detection framework that examines fine-grained token-level interactions across model layers. Our analysis uncovers two characteristic signatures of hallucinated tokens: (i) they yield diffuse, non-localized attention patterns, in contrast to the compact, well-focused attention seen in faithful tokens; and (ii) they fail to exhibit meaningful semantic alignment with any visual region. Guided by these findings, we develop a lightweight and interpretable detection method that leverages patch-level statistical features, combined with hidden-layer representations. Our approach achieves up to 90% accuracy in token-level hallucination detection, demonstrating the superiority of fine-grained structural analysis for detecting hallucinations.

Overthinking Causes Hallucination: Tracing Confounder Propagation in Vision Language Models

Vision Language models (VLMs) often hallucinate non-existent objects. Detecting hallucination is analogous to detecting deception: a single final statement is insufficient, one must examine the underlying reasoning process. Yet existing detectors rely mostly on final-layer signals. Attention-based methods assume hallucinated tokens exhibit low attention, while entropy-based ones use final-step uncertainty. Our analysis reveals the opposite: hallucinated objects can exhibit peaked attention due to contextual priors; and models often express high confidence because intermediate layers have already converged to an incorrect hypothesis. We show that the key to hallucination detection lies within the model's thought process, not its final output. By probing decoder layers, we uncover a previously overlooked behavior, overthinking: models repeatedly revise object hypotheses across layers before committing to an incorrect answer. Once the model latches onto a confounded hypothesis, it can propagate through subsequent layers, ultimately causing hallucination. To capture this behavior, we introduce the Overthinking Score, a metric to measure how many competing hypotheses the model entertains and how unstable these hypotheses are across layers. This score significantly improves hallucination detection: 78.9% F1 on MSCOCO and 71.58% on AMBER.

AstroMLab 4: Benchmark-Topping Performance in Astronomy Q&A with a 70B-Parameter Domain-Specialized Reasoning Model

General-purpose large language models, despite their broad capabilities, often struggle with specialized domain knowledge, a limitation particularly pronounced in more accessible, lower-parameter versions. This gap hinders their deployment as effective agents in demanding fields such as astronomy. Building on our prior work with AstroSage-8B, this study introduces AstroSage-70B, a significantly larger and more advanced domain-specialized natural-language AI assistant. It is designed for research and education across astronomy, astrophysics, space science, astroparticle physics, cosmology, and astronomical instrumentation. Developed from the Llama-3.1-70B foundation, AstroSage-70B underwent extensive continued pre-training on a vast corpus of astronomical literature, followed by supervised fine-tuning and model merging. Beyond its 70-billion parameter scale, this model incorporates refined datasets, judiciously chosen learning hyperparameters, and improved training procedures, achieving state-of-the-art performance on complex astronomical tasks. Notably, we integrated reasoning chains into the SFT dataset, enabling AstroSage-70B to either answer the user query immediately, or first emit a human-readable thought process. Evaluated on the AstroMLab-1 benchmark -- comprising 4,425 questions from literature withheld during training -- AstroSage-70B achieves state-of-the-art performance. It surpasses all other tested open-weight and proprietary models, including leading systems like o3, Gemini-2.5-Pro, Claude-3.7-Sonnet, Deepseek-R1, and Qwen-3-235B, even those with API costs two orders of magnitude higher. This work demonstrates that domain specialization, when applied to large-scale models, can enable them to outperform generalist counterparts in specialized knowledge areas like astronomy, thereby advancing the frontier of AI capabilities in the field.

Empirically evaluating commonsense intelligence in large language models with large-scale human judgments

Commonsense intelligence in machines is often assessed by static benchmarks that compare a model's output against human-prescribed correct labels. An important, albeit implicit, assumption of these labels is that they accurately capture what any human would think, effectively treating human common sense as homogeneous. However, recent empirical work has shown that humans vary enormously in what they consider commonsensical; thus what appears self-evident to one benchmark designer may not be so to another. Here, we propose a novel method for evaluating common sense in artificial intelligence (AI), specifically in large language models (LLMs), that incorporates empirically observed heterogeneity among humans by measuring the correspondence between a model's judgment and that of a human population. We first find that, when treated as independent survey respondents, most LLMs remain below the human median in their individual commonsense competence. Second, when used as simulators of a hypothetical population, LLMs correlate with real humans only modestly in the extent to which they agree on the same set of statements. In both cases, smaller, open-weight models are surprisingly more competitive than larger, proprietary frontier models. Our evaluation framework, which ties commonsense intelligence to its cultural basis, contributes to the growing call for adapting AI models to human collectivities that possess different, often incompatible, social stocks of knowledge.

AstroMLab 2: AstroLLaMA-2-70B Model and Benchmarking Specialised LLMs for Astronomy

Continual pretraining of large language models on domain-specific data has been proposed to enhance performance on downstream tasks. In astronomy, the previous absence of astronomy-focused benchmarks has hindered objective evaluation of these specialized LLM models. Leveraging a recent initiative to curate high-quality astronomical MCQs, this study aims to quantitatively assess specialized LLMs in astronomy. We find that the previously released AstroLLaMA series, based on LLaMA-2-7B, underperforms compared to the base model. We demonstrate that this performance degradation can be partially mitigated by utilizing high-quality data for continual pretraining, such as summarized text from arXiv. Despite the observed catastrophic forgetting in smaller models, our results indicate that continual pretraining on the 70B model can yield significant improvements. However, the current supervised fine-tuning dataset still constrains the performance of instruct models. In conjunction with this study, we introduce a new set of models, AstroLLaMA-3-8B and AstroLLaMA-2-70B, building upon the previous AstroLLaMA series.

AstroMLab 1: Who Wins Astronomy Jeopardy!?

We present a comprehensive evaluation of proprietary and open-weights large language models using the first astronomy-specific benchmarking dataset. This dataset comprises 4,425 multiple-choice questions curated from the Annual Review of Astronomy and Astrophysics, covering a broad range of astrophysical topics. Our analysis examines model performance across various astronomical subfields and assesses response calibration, crucial for potential deployment in research environments. Claude-3.5-Sonnet outperforms competitors by up to 4.6 percentage points, achieving 85.0% accuracy. For proprietary models, we observed a universal reduction in cost every 3-to-12 months to achieve similar score in this particular astronomy benchmark. Open-source models have rapidly improved, with LLaMA-3-70b (80.6%) and Qwen-2-72b (77.7%) now competing with some of the best proprietary models. We identify performance variations across topics, with non-English-focused models generally struggling more in exoplanet-related fields, stellar astrophysics, and instrumentation related questions. These challenges likely stem from less abundant training data, limited historical context, and rapid recent developments in these areas. This pattern is observed across both open-weights and proprietary models, with regional dependencies evident, highlighting the impact of training data diversity on model performance in specialized scientific domains. Top-performing models demonstrate well-calibrated confidence, with correlations above 0.9 between confidence and correctness, though they tend to be slightly underconfident. The development for fast, low-cost inference of open-weights models presents new opportunities for affordable deployment in astronomy. The rapid progress observed suggests that LLM-driven research in astronomy may become feasible in the near future.

Federated PCA on Grassmann Manifold for IoT Anomaly Detection

With the proliferation of the Internet of Things (IoT) and the rising interconnectedness of devices, network security faces significant challenges, especially from anomalous activities. While traditional machine learning-based intrusion detection systems (ML-IDS) effectively employ supervised learning methods, they possess limitations such as the requirement for labeled data and challenges with high dimensionality. Recent unsupervised ML-IDS approaches such as AutoEncoders and Generative Adversarial Networks (GAN) offer alternative solutions but pose challenges in deployment onto resource-constrained IoT devices and in interpretability. To address these concerns, this paper proposes a novel federated unsupervised anomaly detection framework, FedPCA, that leverages Principal Component Analysis (PCA) and the Alternating Directions Method Multipliers (ADMM) to learn common representations of distributed non-i.i.d. datasets. Building on the FedPCA framework, we propose two algorithms, FEDPE in Euclidean space and FEDPG on Grassmann manifolds. Our approach enables real-time threat detection and mitigation at the device level, enhancing network resilience while ensuring privacy. Moreover, the proposed algorithms are accompanied by theoretical convergence rates even under a subsampling scheme, a novel result. Experimental results on the UNSW-NB15 and TON-IoT datasets show that our proposed methods offer performance in anomaly detection comparable to nonlinear baselines, while providing significant improvements in communication and memory efficiency, underscoring their potential for securing IoT networks.

AstroLLaMA-Chat: Scaling AstroLLaMA with Conversational and Diverse Datasets

We explore the potential of enhancing LLM performance in astronomy-focused question-answering through targeted, continual pre-training. By employing a compact 7B-parameter LLaMA-2 model and focusing exclusively on a curated set of astronomy corpora -- comprising abstracts, introductions, and conclusions -- we achieve notable improvements in specialized topic comprehension. While general LLMs like GPT-4 excel in broader question-answering scenarios due to superior reasoning capabilities, our findings suggest that continual pre-training with limited resources can still enhance model performance on specialized topics. Additionally, we present an extension of AstroLLaMA: the fine-tuning of the 7B LLaMA model on a domain-specific conversational dataset, culminating in the release of the chat-enabled AstroLLaMA for community use. Comprehensive quantitative benchmarking is currently in progress and will be detailed in an upcoming full paper. The model, AstroLLaMA-Chat, is now available at https://huggingface.co/universeTBD, providing the first open-source conversational AI tool tailored for the astronomy community.