The paper investigates whether public medical vision‑language benchmarks contain data that may have leaked into the training sets of open VLMs. Using four contamination detectors, the authors find that a significant portion of images in SLAKE‑En have close neighbours in PMC‑OA‑beta, indicating source overlap rather than exact memorization. On the text side, only the exchangeability detector consistently flags potential contamination, while the other methods produce false positives and are unreliable without external baselines.
arXiv:2606. 10066v1 Announce Type: cross Abstract: Medical vision-language models (VLMs) are evaluated on public benchmarks whose images and question-answer pairs have been freely downloadable for years, yet reported accuracy assumes these examples were absent from pretraining.
arXiv:2609.15635v1 Announce Type: cross Abstract: A radiology report can already answer a clinical question, so it is hard to tell whether a vision-language model also uses the image. ModaLens, a pai...
arXiv:2604. 27720v2 Announce Type: replace Abstract: Vision-language models (VLMs) are increasingly applied to medical visual question answering (Med-VQA), yet whether they can \emph{localize} the evidence behind their answers---a prerequisite for clinical auditability---is poorly characterized.
arXiv:2606. 02276v1 Announce Type: cross Abstract: Vision-language models (VLMs) trained on paired chest radiographs and radiology reports learn a shared embedding space that can preserve instance-level image-report correspondence.
A radiology report can already answer a clinical question, so it is hard to tell whether a vision-language model also uses the image. ModaLens, a paired image-swap audit, measures how report availabil...
The study identifies a specific attention head (L8H10) in Vision Transformers that reads annotation text on chest X‑ray images instead of lung pathology, affecting 62% of COVID‑19 images. Causal inpainting experiments confirm that the head’s behavior is driven by annotation content, and removing the head increases lung attention by 1.52×. This spurious annotation‑attending behavior is consistent across multiple ViT architectures and datasets, persists after full fine‑tuning, and is invisible to standard output‑level evaluation, highlighting the need for mechanistic auditing.
arXiv:2606. 17710v1 Announce Type: cross Abstract: Medical vision-language models report strong chest radiograph accuracy, and this is increasingly read as evidence that they use the image.
arXiv:2609.00868v1 Announce Type: cross Abstract: Vision-language models are evaluated by aggregate accuracy on multimodal benchmarks, a practice that implicitly assumes the model uses its visual inp...
The paper investigates how different design choices in visual-head discovery scores affect the attention heads selected in vision‑language models. By systematically varying query type, discovery corpus, and attention source in Qwen3‑VL‑8B‑Instruct, the authors find that the attention source and query type have a larger impact on head selection than the corpus. While necessity scores distinguish discovery settings, steerability does not show a significant advantage over random head sets.
arXiv:2609. 31450v1 Announce Type: cross Abstract: Medical vision-language model (VLM) post-training is commonly evaluated through answer accuracy.
arXiv:2606.15910v3 Announce Type: replace Abstract: An answer from a vision-language model (VLM) to a medical image question can be accepted without review only under a threshold bounding the error o...
The paper investigates how medical vision‑language models (VLMs) behave when faced with distribution shifts such as changes in acquisition domain, supervision, or evaluation protocol. Using datasets like NIH ChestXray14, CheXpert, PadChest, and OpenI, the authors isolate cross‑dataset visual transfer, evaluate multimodal alignment, and quantify source‑proxy leakage in frozen embeddings. They find that self‑supervised visual initialization improves transfer, adversarial adaptation is only marginally helpful, and that multimodal retrieval performance drops under external stress tests while source‑proxy information remains recoverable, highlighting hidden failure modes in medical VLMs.