Paper Detail
Mohammed Safi Ur Rahman Khan, Sanjay Suryanarayanan, Tushar Anand, Mitesh M. Khapra
Large Vision-Language Models (VLMs) are increasingly used to evaluate outputs of other models, for image-to-text (I2T) tasks such as visual question answering, and text-to-image (T2I) generation tasks. Despite this growing reliance, the reliability of these Evaluator VLMs remains under explored. In this work, we systematically evaluate the reliability of Evaluator VLMs across both I2T and T2I tasks. We introduce targeted perturbations that degrade output quality along key error dimensions, including object hallucinations, spatial reasoning, factual grounding, and visual fidelity. These perturbations test whether Evaluator VLMs can reliably account for these quality degrading errors in their evaluations. Using a comprehensive benchmark of over 4000 perturbed instances spanning 40 perturbation dimensions, we evaluate 4 prominent VLMs using single-answer scoring, pairwise comparison, and reference-guided paradigms. Our findings reveal that current VLM evaluators exhibit substantial blind spots: they often fail to detect perturbed outputs - in some cases exceeding 50%, struggle particularly with fine-grained compositional and spatial errors, and are often insensitive to hallucinated content that contradicts the input image. Pairwise comparison proves more reliable, though failure rates persist. These results highlight the unreliable nature of current Evaluator VLMs and urge caution in their deployment for benchmarking and development decisions. Code and data have been made publicly available.
No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.
No ranking explanation is available yet.
No tags.
@misc{khan2026seeing,
title = {Seeing Isn't Believing: Uncovering Blind Spots in Evaluator Vision-Language Models},
author = {Mohammed Safi Ur Rahman Khan and Sanjay Suryanarayanan and Tushar Anand and Mitesh M. Khapra},
year = {2026},
abstract = {Large Vision-Language Models (VLMs) are increasingly used to evaluate outputs of other models, for image-to-text (I2T) tasks such as visual question answering, and text-to-image (T2I) generation tasks. Despite this growing reliance, the reliability of these Evaluator VLMs remains under explored. In this work, we systematically evaluate the reliability of Evaluator VLMs across both I2T and T2I tasks. We introduce targeted perturbations that degrade output quality along key error dimensions, inclu},
url = {https://huggingface.co/papers/2604.21523},
keywords = {vision-language models, image-to-text, text-to-image, visual question answering, hallucinations, spatial reasoning, factual grounding, visual fidelity, perturbations, evaluation reliability, pairwise comparison, benchmarking, code available, huggingface daily},
eprint = {2604.21523},
archiveprefix = {arXiv},
}
{}