Skip to content
Search
paperSeptember 2026Unreviewed

Knowing What Not to Answer: Selective Non-Compliance in Vision-Language Models

Minji Kim, Jihyoung Jang, Hyounghun Kim

View resourcePDF

Abstract

Vision-language models (VLMs) are expected to respond helpfully to appropriate requests while withholding compliance with requests that are incorrect, unsafe, infeasible, or unanswerable. However, existing benchmarks predominantly evaluate non-compliance at the level of the query as a whole, assuming that each request either warrants compliance or requires withholding compliance. In practice, real-world queries can contain a mixture of answerable content and components for which compliance shoul

Categories

Benchmarks & Evaluation

Cite

@misc{kim2026knowing,
  title = {{Knowing What Not to Answer: Selective Non-Compliance in Vision-Language Models}},
  author = {Minji Kim and Jihyoung Jang and Hyounghun Kim},
  year = {2026},
  month = sep,
  eprint = {2609.04720},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.04720}
}