VLM Hallucination Mitigation

Multimodal Large Language Models often suffer from object hallucination. While existing research utilizes attention enhancement and visual retracing, we find these works lack sufficient interpretability regarding attention drift in final model stages. In this paper, we investigate the layer wise evolution of visual features and discover that hallucination stems from deep layer attention regressing toward initial visual noise from early layers. We observe that output reliability depends on acquiring visual anchors at intermediate layers rather than final layers. Based on these insights, we propose CLVA, which stands for Cross-Layer Visual Anchors, a training free method that reinforces critical mid layer features while suppressing regressive noise. This approach effectively pulls deep layer attention back to correct visual regions by utilizing essential anchors captured from attention dynamics. We evaluate our method across diverse architectures and benchmarks, demonstrating outstanding performance without significant increase in computational time and GPU memory.

Object hallucination in Large Vision-Language Models (LVLMs) severely compromises their reliability in real-world applications, posing a critical barrier to their deployment in high-stakes scenarios such as autonomous driving and medical image analysis. Through systematic empirical investigation, we identify that the imbalanced attention allocation, both across modalities (i.e., vision and language) and within modalities (among individual tokens), exhibits a strong causal correlation with the occurrence of object hallucination. Leveraging this insight, we introduce a novel concept termed attention imbalance, which not only quantifies the degree of attention disparity but also visually delineates the underlying patterns (e.g., over-attentiveness to irrelevant language tokens or under-attentiveness to discriminative visual features) that drive object hallucination. To mitigate object hallucination, we further propose Attention Imbalance Rectification (AIR), a lightweight decoding-time intervention method that reallocates attention weights and adjusts attention distributions to rectify modality-wise and token-wise imbalances. Extensive evaluations on four mainstream LVLMs and three benchmarks (CHAIR, POPE, and MM-Vet) with seven baselines demonstrate that AIR consistently reduces object hallucination rates, achieving up to a 35.1% reduction compared to the baselines, while improving up to 15.9% of LVLMs' general capability across diverse vision-language tasks.

Large vision-language models (LVLMs) are increasingly being applied to multi-view image inputs captured from diverse viewpoints. Despite this growing use, current LVLMs often generate incorrect responses due to visual interference from non-target instances or viewpoints, a phenomenon we term multi-view hallucination (MVH). To systematically analyze this problem, we construct MVH-Bench, a benchmark comprising 4.8k question-answer pairs targeting two types of hallucination: cross-instance and cross-view. Empirical results show that MVH is prevalent across recent LVLMs. To address this issue, we propose Reference Shift Contrastive Decoding (RSCD), a training-free decoding technique that suppresses visual interference by generating negative logits through attention masking. Experiments on MVH-Bench with LLaVA-OneVision and Qwen2.5-VL demonstrate that RSCD improves performance by up to 25.7 and 94.8 points over existing hallucination mitigation methods, highlighting the effectiveness of RSCD.

Vision-Language Models (VLMs) have enabled interpretable medical diagnosis by integrating visual perception with linguistic reasoning. Yet, existing medical chain-of-thought (CoT) models lack explicit mechanisms to represent and enforce causal reasoning, leaving them vulnerable to spurious correlations and limiting their clinical reliability. We pinpoint three core challenges in medical CoT reasoning: how to adaptively trigger causal correction, construct high-quality causal-spurious contrastive samples, and maintain causal consistency across reasoning trajectories. To address these challenges, we propose MedCausalX, an end-to-end framework explicitly models causal reasoning chains in medical VLMs. We first introduce the CRMed dataset providing fine-grained anatomical annotations, structured causal reasoning chains, and counterfactual variants that guide the learning of causal relationships beyond superficial correlations. Building upon CRMed, MedCausalX employs a two-stage adaptive reflection architecture equipped with $\langle$causal$\rangle$ and $\langle$verify$\rangle$ tokens, enabling the model to autonomously determine when and how to perform causal analysis and verification. Finally, a trajectory-level causal correction objective optimized through error-attributed reinforcement learning refines the reasoning chain, allowing the model to distinguish genuine causal dependencies from shortcut associations. Extensive experiments on multiple benchmarks show that MedCausalX consistently outperforms state-of-the-art methods, improving diagnostic consistency by +5.4 points, reducing hallucination by over 10 points, and attaining top spatial grounding IoU, thereby setting a new standard for causally grounded medical reasoning. The code and dataset are available at https://github.com/zhcz328/MedCausalX.

Vision-language models (VLMs) enable flexible generative optical character recognition (OCR), while their open-ended decoders can expose wrong but fluent text with weak visual support. In audit-sensitive records, such an output can be more costly than abstention. Frozen or externally served VLMs therefore require an external decision layer that can determine whether a transcription has sufficient visual evidence for release. We introduce the Geometric Risk Controller (GRC), a model-agnostic controller that treats controlled geometric transformations as repeatable black-box probes, screens structurally implausible continuations, and releases the unique candidate supported by coherent cross-view evidence. The protocol provides empirical selective exposure control with explicit coverage and query cost under a reproducible fixed decision rule. Experiments across frozen VLMs and standard scene-text benchmarks consistently reduce mean, upper-tail, and catastrophic error among released outputs while retaining high coverage.

Medical Vision-Language Models (VLMs) often hallucinate by generating responses based on language priors rather than visual evidence, posing risks in clinical applications. We propose Visual Grounding Score Guided Decoding (VGS-Decoding), a training-free method to mitigate hallucinations during inference. Our key insight is that hallucinated tokens maintain or increase their probability when visual information is degraded, while visually grounded tokens decrease in probability. We introduce the Visual Grounding Score (VGS), which measures each token's visual dependency by comparing distributions from original and distorted images. During decoding, we reweight probabilities by amplifying visually grounded tokens while suppressing hallucinations. Unlike fixed-weight contrastive methods, VGS-Decoding provides per-token adaptive control. Experiments on MIMIC-Diff-VQA and VQA-RAD across LLaVA-Med, CheXagent, and MedGemma demonstrate consistent improvements, with up to +9.12% overall gain and $+8.98\%$ in open-ended recall, while introducing only $2\times$ inference overhead and no additional training, making it practical for clinical deployment. Upon acceptance, code will be released publicly to facilitate reproducibility.

Visual-Language Models (VLMs) have achieved remarkable progress in image captioning, visual question answering, and visual reasoning. Yet they remain prone to vision-language misalignment, often producing overly generic or hallucinated descriptions. Existing approaches address this via instruction tuning-requiring costly, large-scale annotated datasets or via complex test-time frameworks for caption refinement. In this work, we revisit image-text alignment through the lens of cycle consistency: given an image and a caption generated by an image-to-text model, the backward mapping through a text-to-image model should reconstruct an image that closely matches the original. In our setup, a VLM serves as the image-to-text component, while a pre-trained text-to-image model closes the loop by reconstructing the image from the generated caption. Building on this, we introduce CycleCap, a fine-tuning scheme to improve image captioning using Group Relative Policy Optimization (GRPO) with a reward based on the similarity between the original and reconstructed images, computed on-the-fly. Unlike previous work that uses cycle consistency loss for preference dataset construction, our method leverages cycle consistency directly as a self-supervised training signal. This enables the use of raw images alone, eliminating the need for curated image-text datasets, while steering the VLM to produce more accurate and grounded text descriptions. Applied to four VLMs ranging from 1B to 7B parameters, CycleCap yields consistent improvements across captioning and hallucination benchmarks, surpassing state-of-the-art methods that rely on supervised cycle consistency training.

Multimodal large language models (MLLMs) struggle with hallucinations, particularly with fine-grained queries, a challenge underrepresented by existing benchmarks that focus on coarse image-related questions. We introduce FIne-grained NEgative queRies (FINER), alongside two benchmarks: FINER-CompreCap and FINER-DOCCI. Using FINER, we analyze hallucinations across four settings: multi-object, multi-attribute, multi-relation, and ``what'' questions. Our benchmarks reveal that MLLMs hallucinate when fine-grained mismatches co-occur with genuinely present elements in the image. To address this, we propose FINER-Tuning, leveraging Direct Preference Optimization (DPO) on FINER-inspired data. Finetuning four frontier MLLMs with FINER-Tuning yields up to 24.2% gains (InternVL3.5-14B) on hallucinations from our benchmarks, while simultaneously improving performance on eight existing hallucination suites, and enhancing general multimodal capabilities across six benchmarks. Code, benchmark, and models are available at https://explainableml.github.io/finer-project/.

Large vision-language models (LVLMs) have become increasingly strong but remain prone to hallucinations in multimodal tasks, which significantly narrows their deployment. As training these LVLMs to avoid hallucinations becomes prohibitively expensive for larger models, training-free methods offer a cheap and flexible solution to this problem, yet existing approaches based on decoding or tool use often bring limited gains and/or weak interpretability. We propose Kestrel, a training-free framework for LVLM hallucination mitigation that combines an explicit visual-grounding agent with evidence-verified self-refinement mechanism. In detail, Kestrel first collects explicit visual evidence and converts tool outputs into reusable and structured textual evidence. Second, to take full advantage of these evidence, Kestrel verifies them via an LVLM judge for evidence checking, then iteratively self-refine answers based on verified evidence to reduce the risk of over-correction. Extensive experiments show that Kestrel improves performance over strong baselines across hallucination benchmarks (e.g., average +3.31% on POPE and +28.34 on MME-Hallucination with Qwen3-VL), while providing transparent verification traces for hallucination diagnosis and analysis -- e.g., both the integrated self-refinement module and grounding agent contributing an average +2.0% gain on POPE.

Large Vision-Language Models (LVLMs) achieve strong performance on many multimodal tasks, but object hallucinations severely undermine their reliability. Most existing studies focus on the text modality, attributing hallucinations to overly strong language priors and insufficient visual grounding. In contrast, we observe that abnormal attention patterns within the visual modality can also give rise to hallucinated objects. Building on this observation, we propose Segmentation-based Attention Entropy (SAE), which leverages semantic segmentation to quantify visual attention uncertainty in a semantic space. Based on SAE, we further design a reliability score for hallucination detection and an SAE-guided attention adjustment method that modifies visual attention at inference time to mitigate hallucinations. We evaluate our approach on public benchmarks and in real embodied multimodal scenarios with quadruped robots. Experiments show that SAE reduces hallucinations without additional training, improving LVLM reliability.

Despite the significant advancements in Large Vision-Language Models (LVLMs), their tendency to generate hallucinations undermines reliability and restricts broader practical deployment. Among the hallucination mitigation methods, feature steering emerges as a promising approach that reduces erroneous outputs in LVLMs without increasing inference costs. However, current methods apply uniform feature steering across all layers. This heuristic strategy ignores inter-layer differences, potentially disrupting layers unrelated to hallucinations and ultimately leading to performance degradation on general tasks. In this paper, we propose Locate-Then-Sparsify for Feature Steering (LTS-FS), a plug-and-play framework which controls the steering intensity according to the hallucination relevance of each layer. We first construct a dataset comprising token-level and sentence-level hallucination cases. Based on this dataset, we introduce an attribution method based on causal interventions to quantify the hallucination relevance of each layer. With the attribution scores across layers, we propose a layerwise strategy that converts these scores into feature steering intensities for individual layers, enabling more precise adjustments specifically on hallucination-relevant layers. Extensive experiments across multiple LVLMs and benchmarks demonstrate that LTS-FS effectively mitigates hallucination while preserving strong performance. Codes are available at https://github.com/huttersadan/LTS-FS.

Recently, Multimodal Large Language Models (MLLMs) have demonstrated significant potential in complex visual tasks through the integration of Chain-of-Thought (CoT) reasoning. However, in Video Question Answering, extended thinking processes do not consistently yield performance gains and may even lead to degradation due to ``visual anchor drifting'', where models increasingly rely on self-generated text, sidelining visual inputs and causing hallucinations. While existing mitigations typically introduce specific mechanisms for the model to re-attend to visual inputs during inference, these approaches often incur prohibitive training costs and suffer from poor generalizability across different architectures. To address this, we propose FrameRepeat, an automated enhancement framework which features a lightweight repeat scoring module that enables Video-LLMs to autonomously identify which frames should be reinforced. We introduce a novel training strategy, Add-One-In (AOI), that uses MLLM output probabilities to generate supervision signals representing repeat gain. This can be used to train a frame scoring network, which guides the frame repetition behavior. Experimental results across multiple models and datasets demonstrate that FrameRepeat is both effective and generalizable in strengthening important visual cues during the reasoning process.

With the growing number and diversity of Vision-Language Models (VLMs), many works explore language-based ensemble, collaboration, and routing techniques across multiple VLMs to improve multi-model reasoning. In contrast, we address the diverse model selection using both vision and language modalities. We introduce focal error diversity to capture complementary reasoning across VLMs and a CKA-based focal diversity metric (CKA-focal) to measure disagreement in their visual embeddings. On the constructed ensemble surface from a pool of candidate VLMs, we applied a Genetic Algorithm to effectively prune out those component VLMs that do not add value to the fusion performance. We identify the best combination for each task as well as fuse the outputs of each VLMs in the model pool, and show that heterogeneous models can capture epistemic uncertainty dynamically and mitigate hallucinations. Our V3Fusion approach is capable of producing dual focal-diversity fused predictions with high performance for vision-language reasoning, even when there is no majority consensus or the majority of VLMs make incorrect predictions. Extensive experiments validate V3Fusion on four popular VLM benchmarks (A-OKVQA, MMMU, MMMU-Pro, and OCR-VQA). The results show that V3Fusion outperforms the best-performing VLM on MMMU by 8.09% and MMMU-Pro by 4.87% gain in accuracy. For generative tasks, V3Fusion outperforms Intern-VL2-8b and Qwen2.5-VL-7b, the top-2 VLM performers on both A-OKVQA and OCR-VQA. Our code and datasets are available at https://github.com/sftekin/v3fusion.

Generating precise diagnostic reports from High-Resolution Computed Tomography (HRCT) is critical for clinical workflow, yet it remains a formidable challenge due to the high pathological diversity and spatial sparsity within 3D volumes. While Video Language Models (VideoLMs) have demonstrated remarkable spatio-temporal reasoning in general domains, their adaptability to domain-specific, high-volume medical interpretation remains underexplored. In this work, we present AbSteering, an abnormality-centric framework that steers VideoLMs toward precise HRCT report generation. Specifically, AbSteering introduces: (i) an abnormality-centric Chain-of-Thought scheme that enforces abnormality reasoning, and (ii) a Direct Preference Optimization objective that utilizes clinically confusable abnormalities as hard negatives to enhance fine-grained discrimination. Our results demonstrate that general-purpose VideoLMs possess strong transferability to high-volume medical imaging when guided by this paradigm. Notably, AbSteering outperforms state-of-the-art domain-specific CT foundation models, which are pretrained with large-scale CTs, achieving superior detection sensitivity while simultaneously mitigating hallucinations.

Vision Language Models (VLMs) exhibit persistent hallucinations in counting tasks, with accuracy substantially lower than other visual reasoning tasks (excluding sentiment). This phenomenon persists even in state-of-the-art reasoning-capable VLMs. Conversely, CNN-based object detection models (ODMs) such as YOLO excel at spatial localization and instance counting with minimal computational overhead. We propose GroundCount, a framework that augments VLMs with explicit spatial grounding from ODMs to mitigate counting hallucinations. In the best case, our prompt-based augmentation strategy achieves 81.3% counting accuracy on the best-performing model (Ovis2.5-2B) - a 6.6pp improvement - while reducing inference time by 22% through elimination of hallucination-driven reasoning loops for stronger models. We conduct comprehensive ablation studies demonstrating that positional encoding is a critical component, being beneficial for stronger models but detrimental for weaker ones. Confidence scores, by contrast, introduce noise for most architectures and their removal improves performance in four of five evaluated models. We further evaluate feature-level fusion architectures, finding that explicit symbolic grounding via structured prompts outperforms implicit feature fusion despite sophisticated cross-attention mechanisms. Our approach yields consistent improvements across four of five evaluated VLM architectures (6.2--7.5pp), with one architecture exhibiting degraded performance due to incompatibility between its iterative reflection mechanisms and structured prompts. These results suggest that counting failures stem from fundamental spatial-semantic integration limitations rather than architecture-specific deficiencies, while highlighting the importance of architectural compatibility in augmentation strategies.

While large vision-language models (LVLMs) achieve strong performance on multimodal tasks, they frequently generate hallucinations -- unfaithful outputs misaligned with the visual input. To address this issue, we introduce CIPHER (Counterfactual Image Perturbations for Hallucination Extraction and Removal), a training-free method that suppresses vision-induced hallucinations via lightweight feature-level correction. Unlike prior training-free approaches that primarily focus on text-induced hallucinations, CIPHER explicitly targets hallucinations arising from the visual modality. CIPHER operates in two phases. In the offline phase, we construct OHC-25K (Object-Hallucinated Counterfactuals, 25,000 samples), a counterfactual dataset consisting of diffusion-edited images that intentionally contradict the original ground-truth captions. We pair these edited images with the unchanged ground-truth captions and process them through an LVLM to extract hallucination-related representations. Contrasting these representations with those from authentic (image, caption) pairs reveals structured, systematic shifts spanning a low-rank subspace characterizing vision-induced hallucination. In the inference phase, CIPHER suppresses hallucinations by projecting intermediate hidden states away from this subspace. Experiments across multiple benchmarks show that CIPHER significantly reduces hallucination rates while preserving task performance, demonstrating the effectiveness of counterfactual visual perturbations for improving LVLM faithfulness. Code and additional materials are available at https://hamidreza-dastmalchi.github.io/cipher-cvpr2026/.

Current training-free methods tackle MLLM hallucination with separate strategies: either enhancing visual signals or suppressing text inertia. However, these separate methods are insufficient due to critical trade-offs: simply enhancing vision often fails against strong language prior, while suppressing language can introduce extra image-irrelevant noise. Moreover, we find their naive combination is also ineffective, necessitating a unified framework. We propose such a framework by focusing on the core asset: the vision token. Our design leverages two key insights: (1) augmented images offer complementary visual semantics, and (2) removing vision tokens (information-gap) isolates hallucination tendencies more precisely than distorting images (modality-gap). Based on these, our framework uses vision tokens in two distinct ways, both operating on latent representations: our Synergistic Visual Calibration (SVC) module incorporates augmented tokens to strengthen visual representations, while our Causal Representation Calibration (CRC) module uses pruned tokens to create latent-space negative samples for correcting internal model biases. By harmonizing these two roles, our framework effectively restores the vision-language balance, significantly reducing object hallucinations, improving POPE accuracy by an average of 2% absolute on LLaVA-1.5 across multiple benchmarks with only a 1.06x inference latency overhead.

Recent advancements in multimodal large reasoning models (MLRMs) have significantly improved performance in visual question answering. However, we observe that transition words (e.g., because, however, and wait) are closely associated with hallucinations and tend to exhibit high-entropy states. We argue that adequate contextual reasoning information can be directly extracted from the token probability distribution. Inspired by superposed representation theory, we propose leveraging latent superposed reasoning to integrate multiple candidate semantics and maintain latent reasoning trajectories. The hypothesis is that reliance on discrete textual inputs may drive the model toward sequential explicit reasoning, underutilizing dense contextual cues during high-entropy reasoning stages. Therefore, we propose constructing rich semantic representations from the token probability distributions to enhance in-context reasoning. With this goal, we present Latent Entropy-Aware Decoding (LEAD), an efficient plug-and-play decoding strategy that leverages semantic context to achieve reliable reasoning. The heart of our method lies in entropy-aware reasoning mode switching. The model employs probability-weighted continuous embeddings under high-entropy states and transitions back to discrete token embeddings as entropy decreases. Moreover, we propose a prior-guided visual anchor injection strategy that encourages the model to focus on visual information. Extensive experiments show that LEAD effectively mitigates hallucinations across various MLRMs on multiple benchmarks.

With the rapid advancement of AIGC technology, developing identification methods to address the security challenges posed by deepfakes has become urgent. Face forgery identification techniques can be categorized into two types: traditional classification methods and explainable VLM approaches. The former provides classification results but lacks explanatory ability, while the latter, although capable of providing coarse-grained explanations, often suffers from hallucinations and insufficient detail. To overcome these limitations, we propose EvolveReason, which mimics the reasoning and observational processes of human auditors when identifying face forgeries. By constructing a chain-of-thought dataset, CoT-Face, tailored for advanced VLMs, our approach guides the model to think in a human-like way, prompting it to output reasoning processes and judgment results. This provides practitioners with reliable analysis and helps alleviate hallucination. Additionally, our framework incorporates a forgery latent-space distribution capture module, enabling EvolveReason to identify high-frequency forgery cues difficult to extract from the original images. To further enhance the reliability of textual explanations, we introduce a self-evolution exploration strategy, leveraging reinforcement learning to allow the model to iteratively explore and optimize its textual descriptions in a two-stage process. Experimental results show that EvolveReason not only outperforms the current state-of-the-art methods in identification performance but also accurately identifies forgery details and demonstrates generalization capabilities.

Although large vision-language models (LVLMs) have demonstrated remarkable capabilities, they are prone to hallucinations in multi-image tasks. We attribute this issue to limitations in existing attention mechanisms and insufficient cross-image modeling. Inspired by this, we propose a structured hallucination mitigation framework involving Cross-Image Attention calibration and Preference Learning (CAPL). CAPL explicitly enhances inter-image interactions at the architectural level while reinforcing reliance on genuine cross-image evidence during training, thereby improving the model's perception and modeling of cross-image associations. Specifically, we (i) introduce a selectable image token interaction attention mechanism to establish fine-grained cross-image entity alignment and information flow; (ii) design a cross-image modeling-based preference optimization strategy that contrasts reasoning outcomes under full inter-image interaction and those obtained when images are mutually invisible, encouraging the model to ground its predictions in authentic visual evidence and mitigating erroneous inferences driven by textual priors. Experimental results demonstrate that CAPL consistently improves performance across multiple model architectures, achieving stable gains on both multi-image hallucination and general benchmarks. Notably, performance on single-image visual tasks remains stable or slightly improves, indicating strong generalization capability.

Hallucination has been a significant impediment to the development and application of current Large Vision-Language Models (LVLMs). To mitigate hallucinations, one intuitive and effective way is to directly increase attention weights to image tokens during inference. Although this effectively reduces the hallucination rate, it often induces repetitive descriptions. To address this, we first conduct an analysis of attention patterns and reveal that real object tokens tend to assign higher attention to the generated text than hallucinated ones. This inspires us to leverage the generated text, which contains instruction-related visual information and contextual knowledge, to alleviate hallucinations while maintaining linguistic coherence. We therefore propose Attention to Generated Text (IAT) and demonstrate that it significantly reduces the hallucination rate while avoiding repetitive descriptions. To prevent naive amplification from impairing the inherent prediction capabilities of LVLMs, we further explore Adaptive IAT (AdaIAT) that employs a layer-wise threshold to control intervention time and fine-grained amplification magnitude tailored to the characteristics of each attention head. Both analysis and experiments demonstrate the effectiveness of AdaIAT. Results of several LVLMs show that AdaIAT effectively alleviates hallucination (reducing hallucination rates $C_S$ and $C_I$ on LLaVA-1.5 by 35.8% and 37.1%, respectively) while preserving linguistic performance and prediction capability, achieving an attractive trade-off.

Multimodal large language models (MLLMs) suffer from pronounced hallucinations in remote sensing visual question-answering (RS-VQA), primarily caused by visual grounding failures in large-scale scenes or misinterpretation of fine-grained small targets. To systematically analyze these issues, we introduce RSHBench, a protocol-based benchmark for fine-grained diagnosis of factual and logical hallucinations. To mitigate grounding-induced factual hallucinations, we further propose Relative Attention-Driven Actively Reasoning (RADAR), a training-free inference method that leverages intrinsic attention in MLLMs to guide progressive localization and fine-grained local reasoning at test time. Extensive experiments across diverse MLLMs demonstrate that RADAR consistently improves RS-VQA performance and reduces both factual and logical hallucinations. Code and data will be publicly available at: https://github.com/MiliLab/RADAR

Reasoning has emerged as a key capability of large language models. In linguistic tasks, this capability can be enhanced by self-improving techniques that refine reasoning paths for subsequent finetuning. However, extending these language-based self-improving approaches to vision language models (VLMs) presents a unique challenge:~visual hallucinations in reasoning paths cannot be effectively verified or rectified. Our solution starts with a key observation about visual contrast: when presented with a contrastive VQA pair, i.e., two visually similar images with synonymous questions, VLMs identify relevant visual cues more precisely. Motivated by this observation, we propose Visual Contrastive Self-Taught Reasoner (VC-STaR), a novel self-improving framework that leverages visual contrast to mitigate hallucinations in model-generated rationales. We collect a diverse suite of VQA datasets, curate contrastive pairs according to multi-modal similarity, and generate rationales using VC-STaR. Consequently, we obtain a new visual reasoning dataset, VisCoR-55K, which is then used to boost the reasoning capability of various VLMs through supervised finetuning. Extensive experiments show that VC-STaR not only outperforms existing self-improving approaches but also surpasses models finetuned on the SoTA visual reasoning datasets, demonstrating that the inherent contrastive ability of VLMs can bootstrap their own visual reasoning. Project at: https://github.com/zhiyupan42/VC-STaR.

Medical Vision-Language Models have shown promising potential in clinical decision support, yet they remain prone to factual hallucinations due to insufficient grounding in localized pathological evidence. Existing medical alignment methods primarily operate at the response level through preference optimization, improving output correctness but leaving intermediate reasoning weakly connected to visual regions. Although chain-of-thought (CoT) enhances multimodal reasoning, it remains largely text-centric, limiting effective integration of clinical visual cues. To address this gap, we propose ClinCoT, a clinical-aware visual chain-of-thought framework that transforms preference optimization from response-level correction to visual-driven reasoning. We introduce an automatic data generation pipeline that constructs clinically grounded preference pairs through reasoning with hypotheses-driven region proposals. Multiple Med-LLMs evaluators rank and assign scores to each response, and these rankings serve as supervision to train the target model. We further introduce a scoring-based margin-aware optimization strategy that incorporates both preference ranking and score difference to refine region-level reasoning trajectories. To maintain alignment as the model's policy evolves during training, we adopt an iterative learning scheme that dynamically regenerates preference data. Extensive experiments on three medical VQA and report generation benchmarks demonstrate that ClinCoT consistently improves factual grounding and achieves superior performance compared with existing preference-based alignment methods.

Reinforcement learning has recently improved the reasoning ability of Large Language Models and Multimodal LLMs, yet prevailing reward designs emphasise final-answer correctness and consequently tolerate process hallucinations--cases where models reach the right answer while misperceiving visual evidence. We address this process-level misalignment with PaLMR, a framework that aligns not only outcomes but also the reasoning process itself. PaLMR comprises two complementary components: a perception-aligned data layer that constructs process-aware reasoning data with structured pseudo-ground-truths and verifiable visual facts, and a process-aligned optimisation layer that constructs a hierarchical reward fusion scheme with a process-aware scoring function to encourage visually faithful chains-of-thought and improve training stability. Experiments on Qwen2.5-VL-7B show that our approach substantially reduces reasoning hallucinations and improves visual reasoning fidelity, achieving state-of-the-art results on HallusionBench while maintaining strong performance on MMMU, MathVista, and MathVerse. These findings indicate that PaLMR offers a principled and practical route to process-aligned multimodal reasoning, advancing the reliability and interpretability of MLLMs.

Although Large Vision-Language Models (LVLMs) have made substantial progress, hallucination, where generated text is not grounded in the visual input, remains a challenge. As LVLMs become stronger, previously reported hallucination patterns, such as linguistic bias and overthinking phenomenon, become far less consistent, making the corresponding mitigation techniques substantially less effective. In this paper, we introduce an Internal self-Correction mechanism utilizing Layer Attention (ICLA) that operates directly on hidden states during generation. Each layer selectively retrieves information from all preceding layers through a diagonal cross-layer attention mechanism, enabling self-refinement without any external correction signals. With introducing and training only 0.2M and 0.1M additional parameters on LLaVA1.5-7B and Qwen2.5-VL-7B, \ours consistently improves visual grounding across multiple hallucination benchmarks, demonstrating its effectiveness for more advanced LVLMs.

Vision-language models (VLMs) show promise in drafting radiology reports, yet they frequently suffer from logical inconsistencies, generating diagnostic impressions unsupported by their own perceptual findings or missing logically entailed conclusions. Standard lexical metrics heavily penalize clinical paraphrasing and fail to capture these deductive failures in reference-free settings. Toward guarantees for clinical reasoning, we introduce a neurosymbolic verification framework that deterministically audits the internal consistency of VLM-generated reports. Our pipeline autoformalizes free-text radiographic findings into structured propositional evidence, utilizing an SMT solver (Z3) and a clinical knowledge base to verify whether each diagnostic claim is mathematically entailed, hallucinated, or omitted. Evaluating seven VLMs across five chest X-ray benchmarks, our verifier exposes distinct reasoning failure modes, such as conservative observation and stochastic hallucination, that remain invisible to traditional metrics. On labeled datasets, enforcing solver-backed entailment acts as a rigorous post-hoc guarantee, systematically eliminating unsupported hallucinations to significantly increase diagnostic soundness and precision in generative clinical assistants.

Multimodal large language models (MLLMs) have achieved remarkable progress in vision-language reasoning, yet they remain vulnerable to hallucination, where generated content deviates from visual evidence. Existing mitigation strategies either require costly supervision during training or introduce additional latency at inference time. Recent vision enhancement methods attempt to address this issue by reinforcing visual tokens during decoding, but they typically inject all tokens indiscriminately, which causes interference from background regions and distracts the model from critical cues. To overcome this challenge, we propose Adaptive Visual Reinforcement (AIR), a training-free framework for MLLMs. AIR consists of two components. Prototype-based token reduction condenses the large pool of visual tokens into a compact subset to suppress redundancy. OT-guided patch reinforcement quantifies the alignment between hidden states and patch embeddings to selectively integrate the most consistent patches into feed-forward layers. As a result, AIR enhances the model's reliance on salient visual information and effectively mitigates hallucination. Extensive experiments across representative MLLMs demonstrate that AIR substantially reduces hallucination while preserving general capabilities, establishing it as an effective solution for building reliable MLLMs.

Automated radiology report generation using vision-language models (VLMs) is limited by the risk of prior-comparison hallucination, where the model generates historical findings unsupported by the current study. We address this challenge with a training-free, inference-time control framework termed Semantically Decoupled Latent Steering (SDLS). Unlike generic activation steering, which often suffers from semantic entanglement, our approach constructs a semantic-free intervention vector via large language model (LLM)-driven semantic decomposition followed by $QR$-based orthogonalization. This orthogonalization step is critical. It leverages geometric constraints to filter out the clinical semantics often entangled in standard principal component analysis (PCA) directions, ensuring that the steering vector targets only the ``historical comparison" axis. We validate our method on the BiomedGPT foundation model, demonstrating that it overcomes the trade-off between hallucination suppression and clinical accuracy. Extensive experiments on MIMIC-CXR, and zero-shot transfer evaluation on CheXpert Plus and IU-Xray, demonstrate the robustness of our approach. Quantitative evaluations on MIMIC-CXR show that our approach significantly reduces the probability of historical hallucinations (FilBERT score decreases from 0.2373 to 0.1889) and improves clinical label fidelity (CheXpert macro-F1 increases from 0.2242 to 0.3208). Supplementary evaluations confirm that the structural integrity of the clinical narrative is maintained.

Chest X-ray plays a central role in thoracic diagnosis, and its interpretation inherently requires multi-step, evidence-grounded reasoning. However, large vision-language models (LVLMs) often generate plausible responses that are not faithfully grounded in diagnostic evidence and provide limited visual evidence for verification, while also requiring costly retraining to support new diagnostic tasks, limiting their reliability and adaptability in clinical settings. To address these limitations, we present CXReasonAgent, a diagnostic agent that integrates a large language model (LLM) with clinically grounded diagnostic tools to perform evidence-grounded diagnostic reasoning using image-derived diagnostic and visual evidence. To evaluate these capabilities, we introduce CXReasonDial, a multi-turn dialogue benchmark with 1,946 dialogues across 12 diagnostic tasks, and show that CXReasonAgent produces faithfully grounded responses, enabling more reliable and verifiable diagnostic reasoning than LVLMs. These findings highlight the importance of integrating clinically grounded diagnostic tools, particularly in safety-critical clinical settings. The demo is available here.